diff --git a/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/.~lock.EstimatedRemainingTimeContext.csv# b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/.~lock.EstimatedRemainingTimeContext.csv#
new file mode 100755
index 0000000000000000000000000000000000000000..e955098031ae47be509270fc5376c7c886105566
--- /dev/null
+++ b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/.~lock.EstimatedRemainingTimeContext.csv#
@@ -0,0 +1 @@
+,DESKTOP-EVD7T0H/diana,DESKTOP-EVD7T0H,06.04.2021 19:22,file:///C:/Users/diana/AppData/Roaming/LibreOffice/4;
\ No newline at end of file
diff --git a/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/EstimatedRemainingTimeContext.csv b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/EstimatedRemainingTimeContext.csv
new file mode 100755
index 0000000000000000000000000000000000000000..1cad682dfa6d442946eaa1f7ce032564607d754d
--- /dev/null
+++ b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/EstimatedRemainingTimeContext.csv
@@ -0,0 +1,141 @@
+name,time,countryCode,ipAddress,level,producer,value
+EstimatedRemainingTimeContext,2021-02-18T11:31:12.752Z,IN,15.206.207.43,3,15.206.207.43,4050
+EstimatedRemainingTimeContext,2021-02-18T11:31:32.752Z,IN,15.206.207.43,3,15.206.207.43,4045.5
+EstimatedRemainingTimeContext,2021-02-18T11:32:02.752Z,IN,15.206.207.43,3,15.206.207.43,4045.5
+EstimatedRemainingTimeContext,2021-02-18T11:32:32.752Z,IN,15.206.207.43,3,15.206.207.43,4045.5
+EstimatedRemainingTimeContext,2021-02-18T11:33:02.752Z,IN,15.206.207.43,3,15.206.207.43,4153.25
+EstimatedRemainingTimeContext,2021-02-18T11:33:32.752Z,IN,15.206.207.43,3,15.206.207.43,4153.25
+EstimatedRemainingTimeContext,2021-02-18T11:34:02.752Z,IN,15.206.207.43,3,15.206.207.43,4153.25
+EstimatedRemainingTimeContext,2021-02-18T11:34:32.752Z,IN,15.206.207.43,3,15.206.207.43,5494.125
+EstimatedRemainingTimeContext,2021-02-18T11:35:02.752Z,IN,15.206.207.43,3,15.206.207.43,5494.125
+EstimatedRemainingTimeContext,2021-02-18T11:35:32.752Z,IN,15.206.207.43,3,15.206.207.43,3080
+EstimatedRemainingTimeContext,2021-02-18T11:36:02.752Z,IN,15.206.207.43,3,15.206.207.43,3080
+EstimatedRemainingTimeContext,2021-02-18T11:36:32.752Z,IN,15.206.207.43,3,15.206.207.43,3080
+EstimatedRemainingTimeContext,2021-02-18T11:37:02.752Z,IN,15.206.207.43,3,15.206.207.43,3348.75
+EstimatedRemainingTimeContext,2021-02-18T11:37:32.752Z,IN,15.206.207.43,3,15.206.207.43,3348.75
+EstimatedRemainingTimeContext,2021-02-18T11:38:02.752Z,IN,15.206.207.43,3,15.206.207.43,4343.625
+EstimatedRemainingTimeContext,2021-02-18T11:38:32.752Z,IN,15.206.207.43,3,15.206.207.43,4394.375
+EstimatedRemainingTimeContext,2021-02-18T11:39:02.752Z,IN,15.206.207.43,3,15.206.207.43,4394.375
+EstimatedRemainingTimeContext,2021-02-18T11:39:32.752Z,IN,15.206.207.43,3,15.206.207.43,4273.5
+EstimatedRemainingTimeContext,2021-02-18T11:40:02.752Z,IN,15.206.207.43,3,15.206.207.43,4268.6875
+EstimatedRemainingTimeContext,2021-02-18T11:40:32.752Z,IN,15.206.207.43,3,15.206.207.43,2845.791666666667
+EstimatedRemainingTimeContext,2021-02-18T11:41:02.752Z,IN,15.206.207.43,3,15.206.207.43,2836.166666666667
+EstimatedRemainingTimeContext,2021-02-18T11:41:32.752Z,IN,15.206.207.43,3,15.206.207.43,2836.166666666667
+EstimatedRemainingTimeContext,2021-02-18T11:42:02.752Z,IN,15.206.207.43,3,15.206.207.43,2836.166666666667
+EstimatedRemainingTimeContext,2021-02-18T11:42:32.752Z,IN,15.206.207.43,3,15.206.207.43,2820.125
+EstimatedRemainingTimeContext,2021-02-18T11:43:02.752Z,IN,15.206.207.43,3,15.206.207.43,2820.125
+EstimatedRemainingTimeContext,2021-02-18T11:43:32.752Z,IN,15.206.207.43,3,15.206.207.43,2813.708333333334
+EstimatedRemainingTimeContext,2021-02-18T11:44:02.752Z,IN,15.206.207.43,3,15.206.207.43,3019.125
+EstimatedRemainingTimeContext,2021-02-18T11:44:32.752Z,IN,15.206.207.43,3,15.206.207.43,2800.875
+EstimatedRemainingTimeContext,2021-02-18T11:45:02.752Z,IN,15.206.207.43,3,15.206.207.43,2788.041666666667
+EstimatedRemainingTimeContext,2021-02-18T11:45:32.752Z,IN,15.206.207.43,3,15.206.207.43,2784.833333333334
+EstimatedRemainingTimeContext,2021-02-18T11:46:02.752Z,IN,15.206.207.43,3,15.206.207.43,2745.5
+EstimatedRemainingTimeContext,2021-02-18T11:46:32.752Z,IN,15.206.207.43,3,15.206.207.43,2736
+EstimatedRemainingTimeContext,2021-02-18T11:47:02.752Z,IN,15.206.207.43,3,15.206.207.43,2049.625
+EstimatedRemainingTimeContext,2021-02-18T11:47:32.752Z,IN,15.206.207.43,3,15.206.207.43,2074.1875
+EstimatedRemainingTimeContext,2021-02-18T11:48:02.752Z,IN,15.206.207.43,3,15.206.207.43,2066.96875
+EstimatedRemainingTimeContext,2021-02-18T11:48:32.752Z,IN,15.206.207.43,3,15.206.207.43,2064.5625
+EstimatedRemainingTimeContext,2021-02-18T11:49:02.752Z,IN,15.206.207.43,3,15.206.207.43,2084.0625
+EstimatedRemainingTimeContext,2021-02-18T11:49:32.752Z,IN,15.206.207.43,3,15.206.207.43,2076.75
+EstimatedRemainingTimeContext,2021-02-18T11:50:02.752Z,IN,15.206.207.43,3,15.206.207.43,2178.125
+EstimatedRemainingTimeContext,2021-02-18T11:50:32.752Z,IN,15.206.207.43,3,15.206.207.43,2186.53125
+EstimatedRemainingTimeContext,2021-02-18T11:51:02.752Z,IN,15.206.207.43,3,15.206.207.43,2183.9375
+EstimatedRemainingTimeContext,2021-02-18T11:51:32.752Z,IN,15.206.207.43,3,15.206.207.43,2123.71875
+EstimatedRemainingTimeContext,2021-02-18T11:52:02.752Z,IN,15.206.207.43,3,15.206.207.43,2116.125
+EstimatedRemainingTimeContext,2021-02-18T11:52:32.752Z,IN,15.206.207.43,3,15.206.207.43,2113.59375
+EstimatedRemainingTimeContext,2021-02-18T11:53:02.752Z,IN,15.206.207.43,3,15.206.207.43,2080
+EstimatedRemainingTimeContext,2021-02-18T11:53:32.752Z,IN,15.206.207.43,3,15.206.207.43,2072.5
+EstimatedRemainingTimeContext,2021-02-18T11:54:02.752Z,IN,15.206.207.43,3,15.206.207.43,2067.5
+EstimatedRemainingTimeContext,2021-02-18T11:54:32.752Z,IN,15.206.207.43,3,15.206.207.43,2060
+EstimatedRemainingTimeContext,2021-02-18T11:55:02.752Z,IN,15.206.207.43,3,15.206.207.43,2055
+EstimatedRemainingTimeContext,2021-02-18T11:55:32.752Z,IN,15.206.207.43,3,15.206.207.43,2050
+EstimatedRemainingTimeContext,2021-02-18T11:56:02.752Z,IN,15.206.207.43,3,15.206.207.43,2007.09375
+EstimatedRemainingTimeContext,2021-02-18T11:56:32.752Z,IN,15.206.207.43,3,15.206.207.43,2004.625
+EstimatedRemainingTimeContext,2021-02-18T11:57:02.752Z,IN,15.206.207.43,3,15.206.207.43,1971.9375
+EstimatedRemainingTimeContext,2021-02-18T11:57:32.752Z,IN,15.206.207.43,3,15.206.207.43,1962.1875
+EstimatedRemainingTimeContext,2021-02-18T11:58:02.752Z,IN,15.206.207.43,3,15.206.207.43,1962.1875
+EstimatedRemainingTimeContext,2021-02-18T11:58:32.752Z,IN,15.206.207.43,3,15.206.207.43,1952.4375
+EstimatedRemainingTimeContext,2021-02-18T11:59:02.752Z,IN,15.206.207.43,3,15.206.207.43,1945.125
+EstimatedRemainingTimeContext,2021-02-18T11:59:32.752Z,IN,15.206.207.43,3,15.206.207.43,1942.6875
+EstimatedRemainingTimeContext,2021-02-18T12:00:02.752Z,IN,15.206.207.43,3,15.206.207.43,1932.9375
+EstimatedRemainingTimeContext,2021-02-18T12:00:32.752Z,IN,15.206.207.43,3,15.206.207.43,1928.0625
+EstimatedRemainingTimeContext,2021-02-18T12:01:02.752Z,IN,15.206.207.43,3,15.206.207.43,1558.275
+EstimatedRemainingTimeContext,2021-02-18T12:01:32.752Z,IN,15.206.207.43,3,15.206.207.43,1483.9
+EstimatedRemainingTimeContext,2021-02-18T12:02:02.752Z,IN,15.206.207.43,3,15.206.207.43,1460.625
+EstimatedRemainingTimeContext,2021-02-18T12:02:32.752Z,IN,15.206.207.43,3,15.206.207.43,1456.875
+EstimatedRemainingTimeContext,2021-02-18T12:03:02.752Z,IN,15.206.207.43,3,15.206.207.43,1464.9
+EstimatedRemainingTimeContext,2021-02-18T12:03:32.752Z,IN,15.206.207.43,3,15.206.207.43,1443.75
+EstimatedRemainingTimeContext,2021-02-18T12:04:02.752Z,IN,15.206.207.43,3,15.206.207.43,1476.475
+EstimatedRemainingTimeContext,2021-02-18T12:04:32.752Z,IN,15.206.207.43,3,15.206.207.43,1504.95
+EstimatedRemainingTimeContext,2021-02-18T12:05:02.752Z,IN,15.206.207.43,3,15.206.207.43,1502.975
+EstimatedRemainingTimeContext,2021-02-18T12:05:32.752Z,IN,15.206.207.43,3,15.206.207.43,1372.4
+EstimatedRemainingTimeContext,2021-02-18T12:06:02.752Z,IN,15.206.207.43,3,15.206.207.43,1365.1
+EstimatedRemainingTimeContext,2021-02-18T12:06:32.752Z,IN,15.206.207.43,3,15.206.207.43,1357.8
+EstimatedRemainingTimeContext,2021-02-18T12:07:02.752Z,IN,15.206.207.43,3,15.206.207.43,1333.8
+EstimatedRemainingTimeContext,2021-02-18T12:07:32.752Z,IN,15.206.207.43,3,15.206.207.43,1346.85
+EstimatedRemainingTimeContext,2021-02-18T12:08:02.752Z,IN,15.206.207.43,3,15.206.207.43,1341.375
+EstimatedRemainingTimeContext,2021-02-18T12:08:32.752Z,IN,15.206.207.43,3,15.206.207.43,1334.075
+EstimatedRemainingTimeContext,2021-02-18T12:09:02.752Z,IN,15.206.207.43,3,15.206.207.43,1299.6
+EstimatedRemainingTimeContext,2021-02-18T12:09:32.752Z,IN,15.206.207.43,3,15.206.207.43,1296
+EstimatedRemainingTimeContext,2021-02-18T12:10:02.752Z,IN,15.206.207.43,3,15.206.207.43,1287
+EstimatedRemainingTimeContext,2021-02-18T12:10:32.752Z,IN,15.206.207.43,3,15.206.207.43,1283.4
+EstimatedRemainingTimeContext,2021-02-18T12:11:02.752Z,IN,15.206.207.43,3,15.206.207.43,1276.2
+EstimatedRemainingTimeContext,2021-02-18T12:11:32.752Z,IN,15.206.207.43,3,15.206.207.43,1059
+EstimatedRemainingTimeContext,2021-02-18T12:12:02.752Z,IN,15.206.207.43,3,15.206.207.43,1054.5
+EstimatedRemainingTimeContext,2021-02-18T12:12:32.752Z,IN,15.206.207.43,3,15.206.207.43,993.3125
+EstimatedRemainingTimeContext,2021-02-18T12:13:02.752Z,IN,15.206.207.43,3,15.206.207.43,990.4375
+EstimatedRemainingTimeContext,2021-02-18T12:13:32.752Z,IN,15.206.207.43,3,15.206.207.43,981.8125
+EstimatedRemainingTimeContext,2021-02-18T12:14:02.752Z,IN,15.206.207.43,3,15.206.207.43,976.0625
+EstimatedRemainingTimeContext,2021-02-18T12:14:32.752Z,IN,15.206.207.43,3,15.206.207.43,973.1875
+EstimatedRemainingTimeContext,2021-02-18T12:15:02.752Z,IN,15.206.207.43,3,15.206.207.43,1036
+EstimatedRemainingTimeContext,2021-02-18T12:15:32.752Z,IN,15.206.207.43,3,15.206.207.43,1003.75
+EstimatedRemainingTimeContext,2021-02-18T12:16:02.752Z,IN,15.206.207.43,3,15.206.207.43,999.1875
+EstimatedRemainingTimeContext,2021-02-18T12:16:32.752Z,IN,15.206.207.43,3,15.206.207.43,978
+EstimatedRemainingTimeContext,2021-02-18T12:17:02.752Z,IN,15.206.207.43,3,15.206.207.43,973.5
+EstimatedRemainingTimeContext,2021-02-18T12:17:32.752Z,IN,15.206.207.43,3,15.206.207.43,830.571428571429
+EstimatedRemainingTimeContext,2021-02-18T12:18:02.752Z,IN,15.206.207.43,3,15.206.207.43,800.017857142857
+EstimatedRemainingTimeContext,2021-02-18T12:18:32.752Z,IN,15.206.207.43,3,15.206.207.43,796.214285714286
+EstimatedRemainingTimeContext,2021-02-18T12:19:02.752Z,IN,15.206.207.43,3,15.206.207.43,802.285714285714
+EstimatedRemainingTimeContext,2021-02-18T12:19:32.752Z,IN,15.206.207.43,3,15.206.207.43,728.357142857143
+EstimatedRemainingTimeContext,2021-02-18T12:20:02.752Z,IN,15.206.207.43,3,15.206.207.43,724.821428571429
+EstimatedRemainingTimeContext,2021-02-18T12:20:32.752Z,IN,15.206.207.43,3,15.206.207.43,696.428571428571
+EstimatedRemainingTimeContext,2021-02-18T12:21:02.752Z,IN,15.206.207.43,3,15.206.207.43,750.571428571429
+EstimatedRemainingTimeContext,2021-02-18T12:21:32.752Z,IN,15.206.207.43,3,15.206.207.43,748.035714285714
+EstimatedRemainingTimeContext,2021-02-18T12:22:02.752Z,IN,15.206.207.43,3,15.206.207.43,741.696428571429
+EstimatedRemainingTimeContext,2021-02-18T12:22:32.752Z,IN,15.206.207.43,3,15.206.207.43,630.125
+EstimatedRemainingTimeContext,2021-02-18T12:23:02.752Z,IN,15.206.207.43,3,15.206.207.43,626.796875
+EstimatedRemainingTimeContext,2021-02-18T12:23:32.752Z,IN,15.206.207.43,3,15.206.207.43,619.03125
+EstimatedRemainingTimeContext,2021-02-18T12:24:02.752Z,IN,15.206.207.43,3,15.206.207.43,615.703125
+EstimatedRemainingTimeContext,2021-02-18T12:24:32.752Z,IN,15.206.207.43,3,15.206.207.43,588.4375
+EstimatedRemainingTimeContext,2021-02-18T12:25:02.752Z,IN,15.206.207.43,3,15.206.207.43,578.59375
+EstimatedRemainingTimeContext,2021-02-18T12:25:32.752Z,IN,15.206.207.43,3,15.206.207.43,576.40625
+EstimatedRemainingTimeContext,2021-02-18T12:26:02.752Z,IN,15.206.207.43,3,15.206.207.43,580.578125
+EstimatedRemainingTimeContext,2021-02-18T12:26:32.752Z,IN,15.206.207.43,3,15.206.207.43,569.171875
+EstimatedRemainingTimeContext,2021-02-18T12:27:02.752Z,IN,15.206.207.43,3,15.206.207.43,568.03125
+EstimatedRemainingTimeContext,2021-02-18T12:27:32.752Z,IN,15.206.207.43,3,15.206.207.43,532.5
+EstimatedRemainingTimeContext,2021-02-18T12:28:02.752Z,IN,15.206.207.43,3,15.206.207.43,520.296875
+EstimatedRemainingTimeContext,2021-02-18T12:28:32.752Z,IN,15.206.207.43,3,15.206.207.43,525.375
+EstimatedRemainingTimeContext,2021-02-18T12:29:02.752Z,IN,15.206.207.43,3,15.206.207.43,477.0625
+EstimatedRemainingTimeContext,2021-02-18T12:29:32.752Z,IN,15.206.207.43,3,15.206.207.43,465.375
+EstimatedRemainingTimeContext,2021-02-18T12:30:02.752Z,IN,15.206.207.43,3,15.206.207.43,463.25
+EstimatedRemainingTimeContext,2021-02-18T12:30:32.752Z,IN,15.206.207.43,3,15.206.207.43,414
+EstimatedRemainingTimeContext,2021-02-18T12:31:02.752Z,IN,15.206.207.43,3,15.206.207.43,407
+EstimatedRemainingTimeContext,2021-02-18T12:31:32.752Z,IN,15.206.207.43,3,15.206.207.43,373.9375
+EstimatedRemainingTimeContext,2021-02-18T12:32:02.752Z,IN,15.206.207.43,3,15.206.207.43,368.125
+EstimatedRemainingTimeContext,2021-02-18T12:32:32.752Z,IN,15.206.207.43,3,15.206.207.43,343.90625
+EstimatedRemainingTimeContext,2021-02-18T12:33:02.752Z,IN,15.206.207.43,3,15.206.207.43,341.96875
+EstimatedRemainingTimeContext,2021-02-18T12:33:32.752Z,IN,15.206.207.43,3,15.206.207.43,330.34375
+EstimatedRemainingTimeContext,2021-02-18T12:34:02.752Z,IN,15.206.207.43,3,15.206.207.43,300.9375
+EstimatedRemainingTimeContext,2021-02-18T12:34:32.752Z,IN,15.206.207.43,3,15.206.207.43,298.125
+EstimatedRemainingTimeContext,2021-02-18T12:35:02.752Z,IN,15.206.207.43,3,15.206.207.43,267.34375
+EstimatedRemainingTimeContext,2021-02-18T12:35:32.752Z,IN,15.206.207.43,3,15.206.207.43,262.734375
+EstimatedRemainingTimeContext,2021-02-18T12:36:02.752Z,IN,15.206.207.43,3,15.206.207.43,234.71875
+EstimatedRemainingTimeContext,2021-02-18T12:36:32.752Z,IN,15.206.207.43,3,15.206.207.43,201.25
+EstimatedRemainingTimeContext,2021-02-18T12:37:02.752Z,IN,15.206.207.43,3,15.206.207.43,194.25
+EstimatedRemainingTimeContext,2021-02-18T12:37:32.752Z,IN,15.206.207.43,3,15.206.207.43,169.296875
+EstimatedRemainingTimeContext,2021-02-18T12:38:02.752Z,IN,15.206.207.43,3,15.206.207.43,142.65625
+EstimatedRemainingTimeContext,2021-02-18T12:38:32.752Z,IN,15.206.207.43,3,15.206.207.43,110.5
+EstimatedRemainingTimeContext,2021-02-18T12:39:02.752Z,IN,15.206.207.43,3,15.206.207.43,60.15625
+EstimatedRemainingTimeContext,2021-02-18T12:39:32.752Z,IN,15.206.207.43,3,15.206.207.43,19.140625
+EstimatedRemainingTimeContext,2021-02-18T12:40:02.752Z,IN,15.206.207.43,3,15.206.207.43,5.359375
+EstimatedRemainingTimeContext,2021-02-18T12:40:32.752Z,IN,15.206.207.43,3,15.206.207.43,0.765625
diff --git a/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/CubeRoot/input_transform.json b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/CubeRoot/input_transform.json
new file mode 100644
index 0000000000000000000000000000000000000000..a8431b20a97f5a993e55b7ca67796039fc808b82
--- /dev/null
+++ b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/CubeRoot/input_transform.json
@@ -0,0 +1 @@
+{"__kind__": "instance", "args": [], "class": "gluonts.transform._base.Chain", "kwargs": {"trans": [{"__kind__": "instance", "args": [], "class": "gluonts.transform.field.RemoveFields", "kwargs": {"field_names": ["feat_dynamic_cat", "feat_static_real", "feat_dynamic_real"]}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.field.SetField", "kwargs": {"output_field": "feat_static_cat", "value": [0.0]}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.field.SetField", "kwargs": {"output_field": "feat_static_real", "value": [0.0]}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.convert.AsNumpyArray", "kwargs": {"dtype": {"__kind__": "type", "class": "numpy.float32"}, "expected_ndim": 1, "field": "feat_static_cat"}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.convert.AsNumpyArray", "kwargs": {"dtype": {"__kind__": "type", "class": "numpy.float32"}, "expected_ndim": 1, "field": "feat_static_real"}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.convert.AsNumpyArray", "kwargs": {"dtype": {"__kind__": "type", "class": "numpy.float32"}, "expected_ndim": 1, "field": "target"}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.feature.AddObservedValuesIndicator", "kwargs": {"dtype": {"__kind__": "type", "class": "numpy.float32"}, "imputation_method": {"__kind__": "instance", "args": [], "class": "gluonts.transform.feature.DummyValueImputation", "kwargs": {"dummy_value": 0.0}}, "output_field": "observed_values", "target_field": "target"}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.feature.AddTimeFeatures", "kwargs": {"output_field": "time_feat", "pred_length": 21, "start_field": "start", "target_field": "target", "time_features": [{"__kind__": "instance", "args": [], "class": "gluonts.time_feature._base.MinuteOfHour", "kwargs": {"normalized": true}}, {"__kind__": "instance", "args": [], "class": "gluonts.time_feature._base.HourOfDay", "kwargs": {"normalized": true}}, {"__kind__": "instance", "args": [], "class": "gluonts.time_feature._base.DayOfWeek", "kwargs": {"normalized": true}}, {"__kind__": "instance", "args": [], "class": "gluonts.time_feature._base.DayOfMonth", "kwargs": {"normalized": true}}, {"__kind__": "instance", "args": [], "class": "gluonts.time_feature._base.DayOfYear", "kwargs": {"normalized": true}}]}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.feature.AddAgeFeature", "kwargs": {"dtype": {"__kind__": "type", "class": "numpy.float32"}, "log_scale": true, "output_field": "feat_dynamic_age", "pred_length": 21, "target_field": "target"}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.convert.VstackFeatures", "kwargs": {"drop_inputs": true, "h_stack": false, "input_fields": ["time_feat", "feat_dynamic_age"], "output_field": "time_feat"}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.split.InstanceSplitter", "kwargs": {"dummy_value": 0.0, "forecast_start_field": "forecast_start", "future_length": 21, "is_pad_field": "is_pad", "lead_time": 0, "output_NTC": true, "past_length": 242, "pick_incomplete": true, "start_field": "start", "target_field": "target", "time_series_fields": ["time_feat", "observed_values"], "train_sampler": {"__kind__": "instance", "args": [], "class": "gluonts.transform.sampler.ExpectedNumInstanceSampler", "kwargs": {"num_instances": 1.0}}}}]}}
diff --git a/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/CubeRoot/parameters.json b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/CubeRoot/parameters.json
new file mode 100644
index 0000000000000000000000000000000000000000..27dabb807b33a28a7e011795cbdb5dbc62271bd6
--- /dev/null
+++ b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/CubeRoot/parameters.json
@@ -0,0 +1 @@
+{"batch_size": 32, "ctx": {"__kind__": "instance", "args": ["cpu", 0], "class": "mxnet.context.Context"}, "dtype": {"__kind__": "type", "class": "numpy.float32"}, "forecast_generator": {"__kind__": "instance", "args": [], "class": "gluonts.model.forecast_generator.SampleForecastGenerator", "kwargs": {}}, "freq": "1min", "input_names": ["feat_static_cat", "feat_static_real", "past_time_feat", "past_target", "past_observed_values", "future_time_feat"], "lead_time": 0, "prediction_length": 21}
diff --git a/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/CubeRoot/prediction_net-0000.params b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/CubeRoot/prediction_net-0000.params
new file mode 100644
index 0000000000000000000000000000000000000000..3b22a424afa379bd3f572a81ff0f500e53f5eb7f
Binary files /dev/null and b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/CubeRoot/prediction_net-0000.params differ
diff --git a/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/CubeRoot/prediction_net-network.json b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/CubeRoot/prediction_net-network.json
new file mode 100644
index 0000000000000000000000000000000000000000..07dc6f139e33ff462a855a5d5c48b88a37e9c88d
--- /dev/null
+++ b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/CubeRoot/prediction_net-network.json
@@ -0,0 +1 @@
+{"__kind__": "instance", "args": [], "class": "gluonts.model.deepar._network.DeepARPredictionNetwork", "kwargs": {"cardinality": [1], "cell_type": "lstm", "context_length": 60, "distr_output": {"__kind__": "instance", "args": [], "class": "gluonts.mx.distribution.student_t.StudentTOutput", "kwargs": {}}, "dropout_rate": 0.1, "dropoutcell_type": "ZoneoutCell", "dtype": {"__kind__": "type", "class": "numpy.float32"}, "embedding_dimension": [1], "history_length": 242, "lags_seq": [1, 2, 3, 4, 5, 6, 7, 58, 59, 60, 61, 62, 118, 119, 120, 121, 122, 178, 179, 180, 181, 182], "num_cells": 40, "num_layers": 2, "num_parallel_samples": 100, "prediction_length": 21, "scaling": true}}
diff --git a/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/CubeRoot/type.txt b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/CubeRoot/type.txt
new file mode 100644
index 0000000000000000000000000000000000000000..6136b11a35db312bcc37697b718d7bcc07495a9e
--- /dev/null
+++ b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/CubeRoot/type.txt
@@ -0,0 +1 @@
+gluonts.mx.model.predictor.RepresentableBlockPredictor
\ No newline at end of file
diff --git a/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/CubeRoot/version.json b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/CubeRoot/version.json
new file mode 100644
index 0000000000000000000000000000000000000000..7866f5ddddc08ed77a6598194d77091373346593
--- /dev/null
+++ b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/CubeRoot/version.json
@@ -0,0 +1 @@
+{"model": "0.6.7", "gluonts": "0.6.7"}
\ No newline at end of file
diff --git a/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/Log/input_transform.json b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/Log/input_transform.json
new file mode 100644
index 0000000000000000000000000000000000000000..a8431b20a97f5a993e55b7ca67796039fc808b82
--- /dev/null
+++ b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/Log/input_transform.json
@@ -0,0 +1 @@
+{"__kind__": "instance", "args": [], "class": "gluonts.transform._base.Chain", "kwargs": {"trans": [{"__kind__": "instance", "args": [], "class": "gluonts.transform.field.RemoveFields", "kwargs": {"field_names": ["feat_dynamic_cat", "feat_static_real", "feat_dynamic_real"]}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.field.SetField", "kwargs": {"output_field": "feat_static_cat", "value": [0.0]}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.field.SetField", "kwargs": {"output_field": "feat_static_real", "value": [0.0]}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.convert.AsNumpyArray", "kwargs": {"dtype": {"__kind__": "type", "class": "numpy.float32"}, "expected_ndim": 1, "field": "feat_static_cat"}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.convert.AsNumpyArray", "kwargs": {"dtype": {"__kind__": "type", "class": "numpy.float32"}, "expected_ndim": 1, "field": "feat_static_real"}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.convert.AsNumpyArray", "kwargs": {"dtype": {"__kind__": "type", "class": "numpy.float32"}, "expected_ndim": 1, "field": "target"}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.feature.AddObservedValuesIndicator", "kwargs": {"dtype": {"__kind__": "type", "class": "numpy.float32"}, "imputation_method": {"__kind__": "instance", "args": [], "class": "gluonts.transform.feature.DummyValueImputation", "kwargs": {"dummy_value": 0.0}}, "output_field": "observed_values", "target_field": "target"}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.feature.AddTimeFeatures", "kwargs": {"output_field": "time_feat", "pred_length": 21, "start_field": "start", "target_field": "target", "time_features": [{"__kind__": "instance", "args": [], "class": "gluonts.time_feature._base.MinuteOfHour", "kwargs": {"normalized": true}}, {"__kind__": "instance", "args": [], "class": "gluonts.time_feature._base.HourOfDay", "kwargs": {"normalized": true}}, {"__kind__": "instance", "args": [], "class": "gluonts.time_feature._base.DayOfWeek", "kwargs": {"normalized": true}}, {"__kind__": "instance", "args": [], "class": "gluonts.time_feature._base.DayOfMonth", "kwargs": {"normalized": true}}, {"__kind__": "instance", "args": [], "class": "gluonts.time_feature._base.DayOfYear", "kwargs": {"normalized": true}}]}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.feature.AddAgeFeature", "kwargs": {"dtype": {"__kind__": "type", "class": "numpy.float32"}, "log_scale": true, "output_field": "feat_dynamic_age", "pred_length": 21, "target_field": "target"}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.convert.VstackFeatures", "kwargs": {"drop_inputs": true, "h_stack": false, "input_fields": ["time_feat", "feat_dynamic_age"], "output_field": "time_feat"}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.split.InstanceSplitter", "kwargs": {"dummy_value": 0.0, "forecast_start_field": "forecast_start", "future_length": 21, "is_pad_field": "is_pad", "lead_time": 0, "output_NTC": true, "past_length": 242, "pick_incomplete": true, "start_field": "start", "target_field": "target", "time_series_fields": ["time_feat", "observed_values"], "train_sampler": {"__kind__": "instance", "args": [], "class": "gluonts.transform.sampler.ExpectedNumInstanceSampler", "kwargs": {"num_instances": 1.0}}}}]}}
diff --git a/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/Log/parameters.json b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/Log/parameters.json
new file mode 100644
index 0000000000000000000000000000000000000000..27dabb807b33a28a7e011795cbdb5dbc62271bd6
--- /dev/null
+++ b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/Log/parameters.json
@@ -0,0 +1 @@
+{"batch_size": 32, "ctx": {"__kind__": "instance", "args": ["cpu", 0], "class": "mxnet.context.Context"}, "dtype": {"__kind__": "type", "class": "numpy.float32"}, "forecast_generator": {"__kind__": "instance", "args": [], "class": "gluonts.model.forecast_generator.SampleForecastGenerator", "kwargs": {}}, "freq": "1min", "input_names": ["feat_static_cat", "feat_static_real", "past_time_feat", "past_target", "past_observed_values", "future_time_feat"], "lead_time": 0, "prediction_length": 21}
diff --git a/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/Log/prediction_net-0000.params b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/Log/prediction_net-0000.params
new file mode 100644
index 0000000000000000000000000000000000000000..f16fac85903e1a26955a2f92b0cc2d1c2494e615
Binary files /dev/null and b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/Log/prediction_net-0000.params differ
diff --git a/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/Log/prediction_net-network.json b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/Log/prediction_net-network.json
new file mode 100644
index 0000000000000000000000000000000000000000..07dc6f139e33ff462a855a5d5c48b88a37e9c88d
--- /dev/null
+++ b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/Log/prediction_net-network.json
@@ -0,0 +1 @@
+{"__kind__": "instance", "args": [], "class": "gluonts.model.deepar._network.DeepARPredictionNetwork", "kwargs": {"cardinality": [1], "cell_type": "lstm", "context_length": 60, "distr_output": {"__kind__": "instance", "args": [], "class": "gluonts.mx.distribution.student_t.StudentTOutput", "kwargs": {}}, "dropout_rate": 0.1, "dropoutcell_type": "ZoneoutCell", "dtype": {"__kind__": "type", "class": "numpy.float32"}, "embedding_dimension": [1], "history_length": 242, "lags_seq": [1, 2, 3, 4, 5, 6, 7, 58, 59, 60, 61, 62, 118, 119, 120, 121, 122, 178, 179, 180, 181, 182], "num_cells": 40, "num_layers": 2, "num_parallel_samples": 100, "prediction_length": 21, "scaling": true}}
diff --git a/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/Log/type.txt b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/Log/type.txt
new file mode 100644
index 0000000000000000000000000000000000000000..6136b11a35db312bcc37697b718d7bcc07495a9e
--- /dev/null
+++ b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/Log/type.txt
@@ -0,0 +1 @@
+gluonts.mx.model.predictor.RepresentableBlockPredictor
\ No newline at end of file
diff --git a/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/Log/version.json b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/Log/version.json
new file mode 100644
index 0000000000000000000000000000000000000000..7866f5ddddc08ed77a6598194d77091373346593
--- /dev/null
+++ b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/Log/version.json
@@ -0,0 +1 @@
+{"model": "0.6.7", "gluonts": "0.6.7"}
\ No newline at end of file
diff --git a/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/NoTransformation/input_transform.json b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/NoTransformation/input_transform.json
new file mode 100644
index 0000000000000000000000000000000000000000..dec43dcab2f4b21c21f2f9f97e083390dcdbf9b4
--- /dev/null
+++ b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/NoTransformation/input_transform.json
@@ -0,0 +1 @@
+{"__kind__": "instance", "args": [], "class": "gluonts.transform._base.Chain", "kwargs": {"trans": [{"__kind__": "instance", "args": [], "class": "gluonts.transform.field.RemoveFields", "kwargs": {"field_names": ["feat_dynamic_cat", "feat_static_real", "feat_dynamic_real"]}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.field.SetField", "kwargs": {"output_field": "feat_static_cat", "value": [0.0]}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.field.SetField", "kwargs": {"output_field": "feat_static_real", "value": [0.0]}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.convert.AsNumpyArray", "kwargs": {"dtype": {"__kind__": "type", "class": "numpy.float32"}, "expected_ndim": 1, "field": "feat_static_cat"}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.convert.AsNumpyArray", "kwargs": {"dtype": {"__kind__": "type", "class": "numpy.float32"}, "expected_ndim": 1, "field": "feat_static_real"}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.convert.AsNumpyArray", "kwargs": {"dtype": {"__kind__": "type", "class": "numpy.float32"}, "expected_ndim": 1, "field": "target"}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.feature.AddObservedValuesIndicator", "kwargs": {"dtype": {"__kind__": "type", "class": "numpy.float32"}, "imputation_method": {"__kind__": "instance", "args": [], "class": "gluonts.transform.feature.DummyValueImputation", "kwargs": {"dummy_value": 0.0}}, "output_field": "observed_values", "target_field": "target"}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.feature.AddTimeFeatures", "kwargs": {"output_field": "time_feat", "pred_length": 21, "start_field": "start", "target_field": "target", "time_features": [{"__kind__": "instance", "args": [], "class": "gluonts.time_feature._base.MinuteOfHour", "kwargs": {"normalized": true}}, {"__kind__": "instance", "args": [], "class": "gluonts.time_feature._base.HourOfDay", "kwargs": {"normalized": true}}, {"__kind__": "instance", "args": [], "class": "gluonts.time_feature._base.DayOfWeek", "kwargs": {"normalized": true}}, {"__kind__": "instance", "args": [], "class": "gluonts.time_feature._base.DayOfMonth", "kwargs": {"normalized": true}}, {"__kind__": "instance", "args": [], "class": "gluonts.time_feature._base.DayOfYear", "kwargs": {"normalized": true}}]}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.feature.AddAgeFeature", "kwargs": {"dtype": {"__kind__": "type", "class": "numpy.float32"}, "log_scale": true, "output_field": "feat_dynamic_age", "pred_length": 21, "target_field": "target"}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.convert.VstackFeatures", "kwargs": {"drop_inputs": true, "h_stack": false, "input_fields": ["time_feat", "feat_dynamic_age"], "output_field": "time_feat"}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.split.InstanceSplitter", "kwargs": {"dummy_value": 0.0, "forecast_start_field": "forecast_start", "future_length": 21, "is_pad_field": "is_pad", "lead_time": 0, "output_NTC": true, "past_length": 232, "pick_incomplete": true, "start_field": "start", "target_field": "target", "time_series_fields": ["time_feat", "observed_values"], "train_sampler": {"__kind__": "instance", "args": [], "class": "gluonts.transform.sampler.ExpectedNumInstanceSampler", "kwargs": {"num_instances": 1.0}}}}]}}
diff --git a/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/NoTransformation/parameters.json b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/NoTransformation/parameters.json
new file mode 100644
index 0000000000000000000000000000000000000000..27dabb807b33a28a7e011795cbdb5dbc62271bd6
--- /dev/null
+++ b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/NoTransformation/parameters.json
@@ -0,0 +1 @@
+{"batch_size": 32, "ctx": {"__kind__": "instance", "args": ["cpu", 0], "class": "mxnet.context.Context"}, "dtype": {"__kind__": "type", "class": "numpy.float32"}, "forecast_generator": {"__kind__": "instance", "args": [], "class": "gluonts.model.forecast_generator.SampleForecastGenerator", "kwargs": {}}, "freq": "1min", "input_names": ["feat_static_cat", "feat_static_real", "past_time_feat", "past_target", "past_observed_values", "future_time_feat"], "lead_time": 0, "prediction_length": 21}
diff --git a/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/NoTransformation/prediction_net-0000.params b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/NoTransformation/prediction_net-0000.params
new file mode 100644
index 0000000000000000000000000000000000000000..7f6523743b2462b6933adb4250853668a6547a52
Binary files /dev/null and b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/NoTransformation/prediction_net-0000.params differ
diff --git a/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/NoTransformation/prediction_net-network.json b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/NoTransformation/prediction_net-network.json
new file mode 100644
index 0000000000000000000000000000000000000000..e452233aee30be1afd5e419e46989f35ecd011b8
--- /dev/null
+++ b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/NoTransformation/prediction_net-network.json
@@ -0,0 +1 @@
+{"__kind__": "instance", "args": [], "class": "gluonts.model.deepar._network.DeepARPredictionNetwork", "kwargs": {"cardinality": [1], "cell_type": "lstm", "context_length": 50, "distr_output": {"__kind__": "instance", "args": [], "class": "gluonts.mx.distribution.student_t.StudentTOutput", "kwargs": {}}, "dropout_rate": 0.1, "dropoutcell_type": "ZoneoutCell", "dtype": {"__kind__": "type", "class": "numpy.float32"}, "embedding_dimension": [1], "history_length": 232, "lags_seq": [1, 2, 3, 4, 5, 6, 7, 58, 59, 60, 61, 62, 118, 119, 120, 121, 122, 178, 179, 180, 181, 182], "num_cells": 40, "num_layers": 2, "num_parallel_samples": 100, "prediction_length": 21, "scaling": true}}
diff --git a/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/NoTransformation/type.txt b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/NoTransformation/type.txt
new file mode 100644
index 0000000000000000000000000000000000000000..6136b11a35db312bcc37697b718d7bcc07495a9e
--- /dev/null
+++ b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/NoTransformation/type.txt
@@ -0,0 +1 @@
+gluonts.mx.model.predictor.RepresentableBlockPredictor
\ No newline at end of file
diff --git a/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/NoTransformation/version.json b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/NoTransformation/version.json
new file mode 100644
index 0000000000000000000000000000000000000000..7866f5ddddc08ed77a6598194d77091373346593
--- /dev/null
+++ b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/NoTransformation/version.json
@@ -0,0 +1 @@
+{"model": "0.6.7", "gluonts": "0.6.7"}
\ No newline at end of file
diff --git a/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/SquareRoot/input_transform.json b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/SquareRoot/input_transform.json
new file mode 100644
index 0000000000000000000000000000000000000000..a8431b20a97f5a993e55b7ca67796039fc808b82
--- /dev/null
+++ b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/SquareRoot/input_transform.json
@@ -0,0 +1 @@
+{"__kind__": "instance", "args": [], "class": "gluonts.transform._base.Chain", "kwargs": {"trans": [{"__kind__": "instance", "args": [], "class": "gluonts.transform.field.RemoveFields", "kwargs": {"field_names": ["feat_dynamic_cat", "feat_static_real", "feat_dynamic_real"]}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.field.SetField", "kwargs": {"output_field": "feat_static_cat", "value": [0.0]}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.field.SetField", "kwargs": {"output_field": "feat_static_real", "value": [0.0]}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.convert.AsNumpyArray", "kwargs": {"dtype": {"__kind__": "type", "class": "numpy.float32"}, "expected_ndim": 1, "field": "feat_static_cat"}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.convert.AsNumpyArray", "kwargs": {"dtype": {"__kind__": "type", "class": "numpy.float32"}, "expected_ndim": 1, "field": "feat_static_real"}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.convert.AsNumpyArray", "kwargs": {"dtype": {"__kind__": "type", "class": "numpy.float32"}, "expected_ndim": 1, "field": "target"}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.feature.AddObservedValuesIndicator", "kwargs": {"dtype": {"__kind__": "type", "class": "numpy.float32"}, "imputation_method": {"__kind__": "instance", "args": [], "class": "gluonts.transform.feature.DummyValueImputation", "kwargs": {"dummy_value": 0.0}}, "output_field": "observed_values", "target_field": "target"}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.feature.AddTimeFeatures", "kwargs": {"output_field": "time_feat", "pred_length": 21, "start_field": "start", "target_field": "target", "time_features": [{"__kind__": "instance", "args": [], "class": "gluonts.time_feature._base.MinuteOfHour", "kwargs": {"normalized": true}}, {"__kind__": "instance", "args": [], "class": "gluonts.time_feature._base.HourOfDay", "kwargs": {"normalized": true}}, {"__kind__": "instance", "args": [], "class": "gluonts.time_feature._base.DayOfWeek", "kwargs": {"normalized": true}}, {"__kind__": "instance", "args": [], "class": "gluonts.time_feature._base.DayOfMonth", "kwargs": {"normalized": true}}, {"__kind__": "instance", "args": [], "class": "gluonts.time_feature._base.DayOfYear", "kwargs": {"normalized": true}}]}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.feature.AddAgeFeature", "kwargs": {"dtype": {"__kind__": "type", "class": "numpy.float32"}, "log_scale": true, "output_field": "feat_dynamic_age", "pred_length": 21, "target_field": "target"}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.convert.VstackFeatures", "kwargs": {"drop_inputs": true, "h_stack": false, "input_fields": ["time_feat", "feat_dynamic_age"], "output_field": "time_feat"}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.split.InstanceSplitter", "kwargs": {"dummy_value": 0.0, "forecast_start_field": "forecast_start", "future_length": 21, "is_pad_field": "is_pad", "lead_time": 0, "output_NTC": true, "past_length": 242, "pick_incomplete": true, "start_field": "start", "target_field": "target", "time_series_fields": ["time_feat", "observed_values"], "train_sampler": {"__kind__": "instance", "args": [], "class": "gluonts.transform.sampler.ExpectedNumInstanceSampler", "kwargs": {"num_instances": 1.0}}}}]}}
diff --git a/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/SquareRoot/parameters.json b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/SquareRoot/parameters.json
new file mode 100644
index 0000000000000000000000000000000000000000..27dabb807b33a28a7e011795cbdb5dbc62271bd6
--- /dev/null
+++ b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/SquareRoot/parameters.json
@@ -0,0 +1 @@
+{"batch_size": 32, "ctx": {"__kind__": "instance", "args": ["cpu", 0], "class": "mxnet.context.Context"}, "dtype": {"__kind__": "type", "class": "numpy.float32"}, "forecast_generator": {"__kind__": "instance", "args": [], "class": "gluonts.model.forecast_generator.SampleForecastGenerator", "kwargs": {}}, "freq": "1min", "input_names": ["feat_static_cat", "feat_static_real", "past_time_feat", "past_target", "past_observed_values", "future_time_feat"], "lead_time": 0, "prediction_length": 21}
diff --git a/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/SquareRoot/prediction_net-0000.params b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/SquareRoot/prediction_net-0000.params
new file mode 100644
index 0000000000000000000000000000000000000000..e4821f6e65a2a0f74b2dc3d5e602d51df617fc01
Binary files /dev/null and b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/SquareRoot/prediction_net-0000.params differ
diff --git a/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/SquareRoot/prediction_net-network.json b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/SquareRoot/prediction_net-network.json
new file mode 100644
index 0000000000000000000000000000000000000000..07dc6f139e33ff462a855a5d5c48b88a37e9c88d
--- /dev/null
+++ b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/SquareRoot/prediction_net-network.json
@@ -0,0 +1 @@
+{"__kind__": "instance", "args": [], "class": "gluonts.model.deepar._network.DeepARPredictionNetwork", "kwargs": {"cardinality": [1], "cell_type": "lstm", "context_length": 60, "distr_output": {"__kind__": "instance", "args": [], "class": "gluonts.mx.distribution.student_t.StudentTOutput", "kwargs": {}}, "dropout_rate": 0.1, "dropoutcell_type": "ZoneoutCell", "dtype": {"__kind__": "type", "class": "numpy.float32"}, "embedding_dimension": [1], "history_length": 242, "lags_seq": [1, 2, 3, 4, 5, 6, 7, 58, 59, 60, 61, 62, 118, 119, 120, 121, 122, 178, 179, 180, 181, 182], "num_cells": 40, "num_layers": 2, "num_parallel_samples": 100, "prediction_length": 21, "scaling": true}}
diff --git a/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/SquareRoot/type.txt b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/SquareRoot/type.txt
new file mode 100644
index 0000000000000000000000000000000000000000..6136b11a35db312bcc37697b718d7bcc07495a9e
--- /dev/null
+++ b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/SquareRoot/type.txt
@@ -0,0 +1 @@
+gluonts.mx.model.predictor.RepresentableBlockPredictor
\ No newline at end of file
diff --git a/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/SquareRoot/version.json b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/SquareRoot/version.json
new file mode 100644
index 0000000000000000000000000000000000000000..7866f5ddddc08ed77a6598194d77091373346593
--- /dev/null
+++ b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/SquareRoot/version.json
@@ -0,0 +1 @@
+{"model": "0.6.7", "gluonts": "0.6.7"}
\ No newline at end of file
diff --git a/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/Standardization/input_transform.json b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/Standardization/input_transform.json
new file mode 100644
index 0000000000000000000000000000000000000000..dec43dcab2f4b21c21f2f9f97e083390dcdbf9b4
--- /dev/null
+++ b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/Standardization/input_transform.json
@@ -0,0 +1 @@
+{"__kind__": "instance", "args": [], "class": "gluonts.transform._base.Chain", "kwargs": {"trans": [{"__kind__": "instance", "args": [], "class": "gluonts.transform.field.RemoveFields", "kwargs": {"field_names": ["feat_dynamic_cat", "feat_static_real", "feat_dynamic_real"]}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.field.SetField", "kwargs": {"output_field": "feat_static_cat", "value": [0.0]}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.field.SetField", "kwargs": {"output_field": "feat_static_real", "value": [0.0]}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.convert.AsNumpyArray", "kwargs": {"dtype": {"__kind__": "type", "class": "numpy.float32"}, "expected_ndim": 1, "field": "feat_static_cat"}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.convert.AsNumpyArray", "kwargs": {"dtype": {"__kind__": "type", "class": "numpy.float32"}, "expected_ndim": 1, "field": "feat_static_real"}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.convert.AsNumpyArray", "kwargs": {"dtype": {"__kind__": "type", "class": "numpy.float32"}, "expected_ndim": 1, "field": "target"}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.feature.AddObservedValuesIndicator", "kwargs": {"dtype": {"__kind__": "type", "class": "numpy.float32"}, "imputation_method": {"__kind__": "instance", "args": [], "class": "gluonts.transform.feature.DummyValueImputation", "kwargs": {"dummy_value": 0.0}}, "output_field": "observed_values", "target_field": "target"}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.feature.AddTimeFeatures", "kwargs": {"output_field": "time_feat", "pred_length": 21, "start_field": "start", "target_field": "target", "time_features": [{"__kind__": "instance", "args": [], "class": "gluonts.time_feature._base.MinuteOfHour", "kwargs": {"normalized": true}}, {"__kind__": "instance", "args": [], "class": "gluonts.time_feature._base.HourOfDay", "kwargs": {"normalized": true}}, {"__kind__": "instance", "args": [], "class": "gluonts.time_feature._base.DayOfWeek", "kwargs": {"normalized": true}}, {"__kind__": "instance", "args": [], "class": "gluonts.time_feature._base.DayOfMonth", "kwargs": {"normalized": true}}, {"__kind__": "instance", "args": [], "class": "gluonts.time_feature._base.DayOfYear", "kwargs": {"normalized": true}}]}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.feature.AddAgeFeature", "kwargs": {"dtype": {"__kind__": "type", "class": "numpy.float32"}, "log_scale": true, "output_field": "feat_dynamic_age", "pred_length": 21, "target_field": "target"}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.convert.VstackFeatures", "kwargs": {"drop_inputs": true, "h_stack": false, "input_fields": ["time_feat", "feat_dynamic_age"], "output_field": "time_feat"}}, {"__kind__": "instance", "args": [], "class": "gluonts.transform.split.InstanceSplitter", "kwargs": {"dummy_value": 0.0, "forecast_start_field": "forecast_start", "future_length": 21, "is_pad_field": "is_pad", "lead_time": 0, "output_NTC": true, "past_length": 232, "pick_incomplete": true, "start_field": "start", "target_field": "target", "time_series_fields": ["time_feat", "observed_values"], "train_sampler": {"__kind__": "instance", "args": [], "class": "gluonts.transform.sampler.ExpectedNumInstanceSampler", "kwargs": {"num_instances": 1.0}}}}]}}
diff --git a/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/Standardization/parameters.json b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/Standardization/parameters.json
new file mode 100644
index 0000000000000000000000000000000000000000..27dabb807b33a28a7e011795cbdb5dbc62271bd6
--- /dev/null
+++ b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/Standardization/parameters.json
@@ -0,0 +1 @@
+{"batch_size": 32, "ctx": {"__kind__": "instance", "args": ["cpu", 0], "class": "mxnet.context.Context"}, "dtype": {"__kind__": "type", "class": "numpy.float32"}, "forecast_generator": {"__kind__": "instance", "args": [], "class": "gluonts.model.forecast_generator.SampleForecastGenerator", "kwargs": {}}, "freq": "1min", "input_names": ["feat_static_cat", "feat_static_real", "past_time_feat", "past_target", "past_observed_values", "future_time_feat"], "lead_time": 0, "prediction_length": 21}
diff --git a/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/Standardization/prediction_net-0000.params b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/Standardization/prediction_net-0000.params
new file mode 100644
index 0000000000000000000000000000000000000000..7d3e2d7c35f78aca852bf80f70b4efc9e9e12294
Binary files /dev/null and b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/Standardization/prediction_net-0000.params differ
diff --git a/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/Standardization/prediction_net-network.json b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/Standardization/prediction_net-network.json
new file mode 100644
index 0000000000000000000000000000000000000000..e452233aee30be1afd5e419e46989f35ecd011b8
--- /dev/null
+++ b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/Standardization/prediction_net-network.json
@@ -0,0 +1 @@
+{"__kind__": "instance", "args": [], "class": "gluonts.model.deepar._network.DeepARPredictionNetwork", "kwargs": {"cardinality": [1], "cell_type": "lstm", "context_length": 50, "distr_output": {"__kind__": "instance", "args": [], "class": "gluonts.mx.distribution.student_t.StudentTOutput", "kwargs": {}}, "dropout_rate": 0.1, "dropoutcell_type": "ZoneoutCell", "dtype": {"__kind__": "type", "class": "numpy.float32"}, "embedding_dimension": [1], "history_length": 232, "lags_seq": [1, 2, 3, 4, 5, 6, 7, 58, 59, 60, 61, 62, 118, 119, 120, 121, 122, 178, 179, 180, 181, 182], "num_cells": 40, "num_layers": 2, "num_parallel_samples": 100, "prediction_length": 21, "scaling": true}}
diff --git a/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/Standardization/type.txt b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/Standardization/type.txt
new file mode 100644
index 0000000000000000000000000000000000000000..6136b11a35db312bcc37697b718d7bcc07495a9e
--- /dev/null
+++ b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/Standardization/type.txt
@@ -0,0 +1 @@
+gluonts.mx.model.predictor.RepresentableBlockPredictor
\ No newline at end of file
diff --git a/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/Standardization/version.json b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/Standardization/version.json
new file mode 100644
index 0000000000000000000000000000000000000000..7866f5ddddc08ed77a6598194d77091373346593
--- /dev/null
+++ b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Models/Standardization/version.json
@@ -0,0 +1 @@
+{"model": "0.6.7", "gluonts": "0.6.7"}
\ No newline at end of file
diff --git a/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Predictions/CubeRoot.csv b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Predictions/CubeRoot.csv
new file mode 100644
index 0000000000000000000000000000000000000000..b531e76577ccd6eeaf469898f36d3937834b576b
--- /dev/null
+++ b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Predictions/CubeRoot.csv
@@ -0,0 +1,43 @@
+original,predictions
+724.821428571429,735.8397461069975
+696.428571428571,670.8208363788767
+750.571428571429,641.8581363424543
+748.035714285714,617.9469646403377
+741.696428571429,594.515870662069
+630.125,595.5019494959064
+626.796875,592.1536742271466
+619.03125,605.3411400633615
+615.703125,567.1618583563142
+588.4375,532.0524877167933
+578.59375,506.3330446400565
+576.40625,507.57373599020065
+580.578125,497.1499216366163
+569.171875,522.8544560239019
+568.03125,531.6509623596843
+532.5,469.5486350232297
+520.296875,502.5019671898981
+525.375,466.1018703255163
+477.0625,461.8723950097421
+465.375,452.73211736138484
+463.25,395.572049746613
+414.0,501.90185075797393
+407.0,518.5398059152202
+373.9375,482.0566473793107
+368.125,458.4988766005031
+343.90625,451.0755536783516
+341.96875,469.21568162604416
+330.34375,466.5797086442084
+300.9375,469.5068989384906
+298.125,458.02597888591475
+267.34375,431.81790182525447
+262.734375,410.0349616597978
+234.71875,444.05571084688034
+201.25,452.42949808993103
+194.25,446.1724466610581
+169.296875,456.15982508091446
+142.65625,461.3549024195875
+110.5,455.3486818131098
+60.15625,477.9184406368721
+19.140625,445.5313873078685
+5.359375,430.44168354915143
+0.765625,443.89768438144586
diff --git a/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Predictions/Log.csv b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Predictions/Log.csv
new file mode 100644
index 0000000000000000000000000000000000000000..e8d6460cc5b10c6a111dbbe1def53f0af411a4e0
--- /dev/null
+++ b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Predictions/Log.csv
@@ -0,0 +1,43 @@
+original,predictions
+724.821428571429,732.63330078125
+696.428571428571,692.6189575195312
+750.571428571429,708.2552490234375
+748.035714285714,717.3744506835938
+741.696428571429,646.7852783203125
+630.125,699.251708984375
+626.796875,669.52294921875
+619.03125,653.7056884765625
+615.703125,655.6961669921875
+588.4375,635.6104125976562
+578.59375,650.6466064453125
+576.40625,602.1397705078125
+580.578125,621.825439453125
+569.171875,622.3018188476562
+568.03125,590.0972290039062
+532.5,608.3982543945312
+520.296875,563.3687133789062
+525.375,590.8246459960938
+477.0625,620.4140625
+465.375,547.7076416015625
+463.25,476.70721435546875
+414.0,562.6091918945312
+407.0,507.80938720703125
+373.9375,507.58282470703125
+368.125,421.6654052734375
+343.90625,385.6040954589844
+341.96875,384.8360290527344
+330.34375,384.41729736328125
+300.9375,366.4085998535156
+298.125,340.28277587890625
+267.34375,339.543701171875
+262.734375,330.5943603515625
+234.71875,335.1805419921875
+201.25,331.268310546875
+194.25,340.6683349609375
+169.296875,341.9048767089844
+142.65625,332.6848449707031
+110.5,317.31134033203125
+60.15625,320.60076904296875
+19.140625,328.8138732910156
+5.359375,325.4771423339844
+0.765625,325.7332458496094
diff --git a/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Predictions/NoTransformation.csv b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Predictions/NoTransformation.csv
new file mode 100644
index 0000000000000000000000000000000000000000..9330ec0f2b8f57abdba4bede4f36923de90b40bb
--- /dev/null
+++ b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Predictions/NoTransformation.csv
@@ -0,0 +1,43 @@
+original,predictions
+724.821428571429,733.8521728515625
+696.428571428571,689.3428955078125
+750.571428571429,670.15185546875
+748.035714285714,622.4755249023438
+741.696428571429,637.7252807617188
+630.125,639.6415405273438
+626.796875,598.9541625976562
+619.03125,581.1470947265625
+615.703125,622.4326782226562
+588.4375,579.5543823242188
+578.59375,544.7575073242188
+576.40625,611.096435546875
+580.578125,588.2201538085938
+569.171875,568.5015869140625
+568.03125,565.6257934570312
+532.5,583.0615234375
+520.296875,534.1814575195312
+525.375,530.15380859375
+477.0625,525.5071411132812
+465.375,517.7937622070312
+463.25,629.8885498046875
+414.0,796.7608032226562
+407.0,896.6124267578125
+373.9375,852.8779907226562
+368.125,825.1931762695312
+343.90625,809.500732421875
+341.96875,781.916015625
+330.34375,741.1702880859375
+300.9375,777.7379760742188
+298.125,848.84814453125
+267.34375,712.3903198242188
+262.734375,701.8955688476562
+234.71875,743.140625
+201.25,832.5770874023438
+194.25,825.59814453125
+169.296875,853.6260375976562
+142.65625,865.0538940429688
+110.5,822.901611328125
+60.15625,809.9740600585938
+19.140625,850.9570922851562
+5.359375,809.3712768554688
+0.765625,819.896728515625
diff --git a/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Predictions/SquareRoot.csv b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Predictions/SquareRoot.csv
new file mode 100644
index 0000000000000000000000000000000000000000..eb8c7277b1f59407aa42628d2712b3ace080c015
--- /dev/null
+++ b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Predictions/SquareRoot.csv
@@ -0,0 +1,43 @@
+original,predictions
+724.821428571429,736.0545654296875
+696.428571428571,721.4066772460938
+750.571428571429,702.3803100585938
+748.035714285714,714.0524291992188
+741.696428571429,658.274658203125
+630.125,645.140869140625
+626.796875,644.2606201171875
+619.03125,655.6788940429688
+615.703125,599.9710083007812
+588.4375,629.6721801757812
+578.59375,582.6422119140625
+576.40625,578.9059448242188
+580.578125,571.5673217773438
+569.171875,639.6375122070312
+568.03125,571.3453979492188
+532.5,549.3250732421875
+520.296875,546.5228271484375
+525.375,496.8221740722656
+477.0625,524.48583984375
+465.375,510.9547424316406
+463.25,569.145751953125
+414.0,714.9201049804688
+407.0,838.8485717773438
+373.9375,1063.24072265625
+368.125,1308.258544921875
+343.90625,1458.994384765625
+341.96875,1575.776611328125
+330.34375,1628.0306396484375
+300.9375,1704.641357421875
+298.125,1785.0584716796875
+267.34375,1796.82861328125
+262.734375,1822.2037353515625
+234.71875,1856.30126953125
+201.25,1915.1636962890625
+194.25,1905.45458984375
+169.296875,1973.171142578125
+142.65625,2016.01318359375
+110.5,2104.3544921875
+60.15625,2119.7578125
+19.140625,2150.51708984375
+5.359375,2172.843017578125
+0.765625,2138.708984375
diff --git a/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Predictions/Standardization.csv b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Predictions/Standardization.csv
new file mode 100644
index 0000000000000000000000000000000000000000..263b10fce42df28a029793ebe68539c279b20663
--- /dev/null
+++ b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/Predictions/Standardization.csv
@@ -0,0 +1,43 @@
+original,predictions
+724.821428571429,756.6295166015625
+696.428571428571,800.9555053710938
+750.571428571429,745.2360229492188
+748.035714285714,778.6434326171875
+741.696428571429,802.8636474609375
+630.125,805.2674560546875
+626.796875,799.1098022460938
+619.03125,890.4715576171875
+615.703125,774.0389404296875
+588.4375,866.7249755859375
+578.59375,846.4957885742188
+576.40625,804.7560424804688
+580.578125,870.5142822265625
+569.171875,799.7620849609375
+568.03125,868.6627197265625
+532.5,867.8150634765625
+520.296875,867.6124267578125
+525.375,888.9931640625
+477.0625,907.0816650390625
+465.375,853.5992431640625
+463.25,884.44970703125
+414.0,696.6505737304688
+407.0,746.9124755859375
+373.9375,804.1859130859375
+368.125,916.2717895507812
+343.90625,1040.192138671875
+341.96875,1260.3509521484375
+330.34375,1379.0322265625
+300.9375,1363.556884765625
+298.125,1550.8292236328125
+267.34375,1460.5450439453125
+262.734375,1617.5673828125
+234.71875,1819.9097900390625
+201.25,1926.43359375
+194.25,1877.6903076171875
+169.296875,1989.887451171875
+142.65625,1952.82373046875
+110.5,2000.62646484375
+60.15625,1992.292236328125
+19.140625,2054.389892578125
+5.359375,2030.25341796875
+0.765625,2182.593505859375
diff --git a/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/cuberoot.ipynb b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/cuberoot.ipynb
new file mode 100644
index 0000000000000000000000000000000000000000..a0433cbf78e9d9ae2ff4fe36eaf67ae839feb442
--- /dev/null
+++ b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/cuberoot.ipynb
@@ -0,0 +1,1947 @@
+{
+ "cells": [
+ {
+ "cell_type": "code",
+ "execution_count": 1,
+ "id": "relative-spanish",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "%matplotlib inline\n",
+ "import mxnet as mx\n",
+ "from mxnet import gluon\n",
+ "import numpy as np\n",
+ "import pandas as pd\n",
+ "import matplotlib.pyplot as plt\n",
+ "import json\n",
+ "from gluonts.dataset.common import ListDataset\n",
+ "from datetime import datetime, timedelta\n",
+ "from gluonts.model.simple_feedforward import SimpleFeedForwardEstimator\n",
+ "from gluonts.model.deepar import DeepAREstimator\n",
+ "from gluonts.mx.trainer import Trainer\n",
+ "from gluonts.evaluation.backtest import make_evaluation_predictions\n",
+ "from gluonts.evaluation import Evaluator\n",
+ "from gluonts.dataset.util import to_pandas\n",
+ "from gluonts.dataset.field_names import FieldName\n",
+ "from gluonts.model.forecast import SampleForecast\n",
+ "from pandas import Timestamp\n",
+ "import itertools\n",
+ "from gluonts.model.estimator import GluonEstimator"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "id": "consolidated-russell",
+ "metadata": {},
+ "source": [
+ "# Data Loading"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 2,
+ "id": "handled-ballet",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "filename='EstimatedRemainingTimeContext'"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 3,
+ "id": "anonymous-evans",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "dataset= pd.read_csv(filename + \".csv\")\n",
+ "X= dataset[['name','time', 'countryCode', 'ipAddress', 'level',\n",
+ " 'producer']]\n",
+ "y= dataset.iloc[:,6]\n",
+ "#changing the names of the attributes\n",
+ "gluon_dataset= pd.DataFrame()\n",
+ "gluon_dataset['ds'] = pd.to_datetime(X[\"time\"])\n",
+ "gluon_dataset['y']=y\n",
+ "gluon_dataset['ds'] = gluon_dataset['ds'].dt.tz_convert(None)\n",
+ "gluon_dataset_copy=gluon_dataset.copy()"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 4,
+ "id": "original-finnish",
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/html": [
+ "\n",
+ "\n",
+ "
\n",
+ " \n",
+ " \n",
+ " | \n",
+ " ds | \n",
+ " y | \n",
+ "
\n",
+ " \n",
+ " \n",
+ " \n",
+ " | 0 | \n",
+ " 2021-02-18 11:31:12.752 | \n",
+ " 4050.000000 | \n",
+ "
\n",
+ " \n",
+ " | 1 | \n",
+ " 2021-02-18 11:31:32.752 | \n",
+ " 4045.500000 | \n",
+ "
\n",
+ " \n",
+ " | 2 | \n",
+ " 2021-02-18 11:32:02.752 | \n",
+ " 4045.500000 | \n",
+ "
\n",
+ " \n",
+ " | 3 | \n",
+ " 2021-02-18 11:32:32.752 | \n",
+ " 4045.500000 | \n",
+ "
\n",
+ " \n",
+ " | 4 | \n",
+ " 2021-02-18 11:33:02.752 | \n",
+ " 4153.250000 | \n",
+ "
\n",
+ " \n",
+ " | ... | \n",
+ " ... | \n",
+ " ... | \n",
+ "
\n",
+ " \n",
+ " | 135 | \n",
+ " 2021-02-18 12:38:32.752 | \n",
+ " 110.500000 | \n",
+ "
\n",
+ " \n",
+ " | 136 | \n",
+ " 2021-02-18 12:39:02.752 | \n",
+ " 60.156250 | \n",
+ "
\n",
+ " \n",
+ " | 137 | \n",
+ " 2021-02-18 12:39:32.752 | \n",
+ " 19.140625 | \n",
+ "
\n",
+ " \n",
+ " | 138 | \n",
+ " 2021-02-18 12:40:02.752 | \n",
+ " 5.359375 | \n",
+ "
\n",
+ " \n",
+ " | 139 | \n",
+ " 2021-02-18 12:40:32.752 | \n",
+ " 0.765625 | \n",
+ "
\n",
+ " \n",
+ "
\n",
+ "
140 rows × 2 columns
\n",
+ "
"
+ ]
+ },
+ "metadata": {
+ "needs_background": "light"
+ },
+ "output_type": "display_data"
+ },
+ {
+ "data": {
+ "image/png": "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\n",
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {
+ "needs_background": "light"
+ },
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "gluon_dataset_copy['y_orig']=gluon_dataset_copy['y']\n",
+ "transformed = cube_root_transformation(gluon_dataset_copy['y'])\n",
+ "gluon_dataset_copy['y']=transformed\n",
+ "\n",
+ "#plot before transformation\n",
+ "pd.plotting.register_matplotlib_converters()\n",
+ "f, ax = plt.subplots(figsize=(15,7))\n",
+ "gluon_dataset_copy.plot(kind='line', x='ds', y='y_orig', color='blue', label='Train', ax=ax)\n",
+ "plt.title(filename)\n",
+ "plt.show()\n",
+ "\n",
+ "#plot after transformation\n",
+ "pd.plotting.register_matplotlib_converters()\n",
+ "f, ax = plt.subplots(figsize=(15,7))\n",
+ "gluon_dataset_copy.plot(kind='line', x='ds', y='y', color='blue', label='Train', ax=ax)\n",
+ "#test.plot(kind='line', x='ds', y='y', color='red', label='Test', ax=ax)\n",
+ "plt.title(filename + \"_transformed using Square Root Transformation\")\n",
+ "plt.show()"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 8,
+ "id": "historic-judgment",
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/html": [
+ "\n",
+ "\n",
+ "
\n",
+ " \n",
+ " \n",
+ " | \n",
+ " ds | \n",
+ " y | \n",
+ " y_orig | \n",
+ "
\n",
+ " \n",
+ " \n",
+ " \n",
+ " | 0 | \n",
+ " 2021-02-18 11:31:12.752 | \n",
+ " 15.939879 | \n",
+ " 4050.000000 | \n",
+ "
\n",
+ " \n",
+ " | 1 | \n",
+ " 2021-02-18 11:32:12.752 | \n",
+ " 15.933973 | \n",
+ " 4045.500000 | \n",
+ "
\n",
+ " \n",
+ " | 2 | \n",
+ " 2021-02-18 11:33:12.752 | \n",
+ " 15.933973 | \n",
+ " 4045.500000 | \n",
+ "
\n",
+ " \n",
+ " | 3 | \n",
+ " 2021-02-18 11:34:12.752 | \n",
+ " 15.933973 | \n",
+ " 4045.500000 | \n",
+ "
\n",
+ " \n",
+ " | 4 | \n",
+ " 2021-02-18 11:35:12.752 | \n",
+ " 16.074200 | \n",
+ " 4153.250000 | \n",
+ "
\n",
+ " \n",
+ " | ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ "
\n",
+ " \n",
+ " | 135 | \n",
+ " 2021-02-18 13:46:12.752 | \n",
+ " 4.798669 | \n",
+ " 110.500000 | \n",
+ "
\n",
+ " \n",
+ " | 136 | \n",
+ " 2021-02-18 13:47:12.752 | \n",
+ " 3.918263 | \n",
+ " 60.156250 | \n",
+ "
\n",
+ " \n",
+ " | 137 | \n",
+ " 2021-02-18 13:48:12.752 | \n",
+ " 2.674969 | \n",
+ " 19.140625 | \n",
+ "
\n",
+ " \n",
+ " | 138 | \n",
+ " 2021-02-18 13:49:12.752 | \n",
+ " 1.750000 | \n",
+ " 5.359375 | \n",
+ "
\n",
+ " \n",
+ " | 139 | \n",
+ " 2021-02-18 13:50:12.752 | \n",
+ " 0.914826 | \n",
+ " 0.765625 | \n",
+ "
\n",
+ " \n",
+ "
\n",
+ "
140 rows × 3 columns
\n",
+ "
"
+ ]
+ },
+ "metadata": {
+ "needs_background": "light"
+ },
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "train_entry = next(iter(train_ds))\n",
+ "train_entry.keys()\n",
+ "\n",
+ "validation_entry = next(iter(validation_ds))\n",
+ "validation_entry.keys()\n",
+ "\n",
+ "test_entry = next(iter(test_ds))\n",
+ "test_entry.keys()\n",
+ "\n",
+ "test_orig_entry = next(iter(test_orig_ds))\n",
+ "test_orig_entry.keys()\n",
+ "\n",
+ "test_series = to_pandas(test_entry)\n",
+ "train_series = to_pandas(train_entry)\n",
+ "validation_series = to_pandas(validation_entry)\n",
+ "\n",
+ "prediction_length = len(test_series) - len(validation_series)\n",
+ "context_length = [50,60]\n",
+ "\n",
+ "\n",
+ "list1=train_series.index\n",
+ "list2=train_entry['target']\n",
+ "\n",
+ "d = {'ds':list1,'y':list2}\n",
+ "df_train=pd.DataFrame(d, columns=['ds','y'])\n",
+ "\n",
+ "fig, ax = plt.subplots( figsize=(10, 7))\n",
+ "\n",
+ "# train_series.plot(ax=ax[0])\n",
+ "# ax[0].grid(which=\"both\")\n",
+ "# ax[0].legend([\"train series\"], loc=\"upper left\")\n",
+ "\n",
+ "test_series.plot(ax=ax)\n",
+ "plt.axvline(train_series.index[-1], color='r') # end of train dataset\n",
+ "plt.axvline(validation_series.index[-1], color='b') # end of validation dataset\n",
+ "ax.grid(which=\"both\")\n",
+ "ax.legend([\"test series\", \"end of train series\",\"end of validation series\"], loc=\"upper left\")\n",
+ "\n",
+ "plt.show()"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 12,
+ "id": "dedicated-might",
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "21"
+ ]
+ },
+ "execution_count": 12,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "prediction_length"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "id": "educational-samoa",
+ "metadata": {},
+ "source": [
+ "# Hyper-parameter Tuning"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 13,
+ "id": "affiliated-journalist",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "batch_size = [75,100]\n",
+ "epochs = [5,10]\n",
+ "num_batches_per_epoch = [10,20]\n",
+ "learning_rate = [1e-3,1e-4]"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 14,
+ "id": "annual-filing",
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "Total Possible Models 32\n"
+ ]
+ }
+ ],
+ "source": [
+ "from sklearn.model_selection import ParameterGrid\n",
+ "\n",
+ "param_grid = {'batch_size': batch_size,\n",
+ " 'epochs': epochs,\n",
+ " 'num_batches_per_epoch': num_batches_per_epoch,\n",
+ " 'learning_rate': learning_rate,\n",
+ " 'context_length': context_length\n",
+ " }\n",
+ "grid = ParameterGrid(param_grid)\n",
+ "cnt = 0\n",
+ "for p in grid:\n",
+ " cnt = cnt+1\n",
+ "\n",
+ "print('Total Possible Models',cnt)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 15,
+ "id": "consolidated-thinking",
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stderr",
+ "output_type": "stream",
+ "text": [
+ " 0%| | 0/10 [00:00"
+ ]
+ },
+ "metadata": {
+ "needs_background": "light"
+ },
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "def plot_prob_forecasts(ts_entry, forecast_entry):\n",
+ " plot_length = len(ts_entry)\n",
+ " prediction_intervals = (50.0,98.0)\n",
+ " legend = [\"observations\", \"median prediction\"] + [f\"{k}% prediction interval\" for k in prediction_intervals][::-1]\n",
+ "\n",
+ " fig, ax = plt.subplots(1, 1, figsize=(10, 7))\n",
+ " ts_entry[-plot_length:].plot(ax=ax) # plot the time series\n",
+ " forecast_entry.plot(prediction_intervals=prediction_intervals, color='g')\n",
+ " plt.grid(which=\"both\")\n",
+ " plt.legend(legend, loc=\"upper left\")\n",
+ " plt.show()\n",
+ "plot_prob_forecasts(gluon_dataset['y'], forecast_entry_inv)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 22,
+ "id": "tight-stranger",
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stderr",
+ "output_type": "stream",
+ "text": [
+ "Running evaluation: 100%|██████████| 1/1 [00:00<00:00, 4.67it/s]\n"
+ ]
+ }
+ ],
+ "source": [
+ "evaluator = Evaluator(quantiles=[0.1, 0.5, 0.9])\n",
+ "agg_metrics, item_metrics = evaluator(iter(tss_inv), iter([forecast_entry_inv]), num_series=len(test_ds))\n"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 23,
+ "id": "hydraulic-biology",
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "{'MSE': 66454.697796436,\n",
+ " 'abs_error': 4667.486553423428,\n",
+ " 'abs_target_sum': 4846.515611285734,\n",
+ " 'abs_target_mean': 230.78645768027303,\n",
+ " 'seasonal_error': 86.9523304104521,\n",
+ " 'MASE': 2.556127747233072,\n",
+ " 'MAPE': 32.82394755220077,\n",
+ " 'sMAPE': 0.7664386009802475,\n",
+ " 'OWA': nan,\n",
+ " 'MSIS': 38.94625974320382,\n",
+ " 'QuantileLoss[0.1]': 5372.805501363753,\n",
+ " 'Coverage[0.1]': 0.9523809523809523,\n",
+ " 'QuantileLoss[0.5]': 4667.486553423428,\n",
+ " 'Coverage[0.5]': 1.0,\n",
+ " 'QuantileLoss[0.9]': 1340.740606846174,\n",
+ " 'Coverage[0.9]': 1.0,\n",
+ " 'RMSE': 257.788086994795,\n",
+ " 'NRMSE': 1.116998326443961,\n",
+ " 'ND': 0.9630602535468132,\n",
+ " 'wQuantileLoss[0.1]': 1.1085913947852526,\n",
+ " 'wQuantileLoss[0.5]': 0.9630602535468132,\n",
+ " 'wQuantileLoss[0.9]': 0.2766401089731533,\n",
+ " 'mean_absolute_QuantileLoss': 3793.6775538777842,\n",
+ " 'mean_wQuantileLoss': 0.7827639191017397,\n",
+ " 'MAE_Coverage': 0.4841269841269842}"
+ ]
+ },
+ "execution_count": 23,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "agg_metrics"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "id": "comparable-newcastle",
+ "metadata": {},
+ "source": [
+ "# Residual"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 24,
+ "id": "protecting-alias",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "#get the mean for all the samples\n",
+ "f1 = forecast_entry_val_inv.samples.mean(axis=0)\n",
+ "f2 = forecast_entry_inv.samples.mean(axis=0)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 25,
+ "id": "stock-representation",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "original_values=list(gluon_dataset['y'])\n",
+ "orig_test = original_values[:len(original_values)-prediction_length]\n",
+ "test_extract=original_values[-prediction_length:]\n",
+ "val_extract=orig_test[-prediction_length:]"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 26,
+ "id": "hourly-philosophy",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "valandtest_orig=list()\n",
+ "valandtest_pred=list()\n",
+ "for a1 in val_extract:\n",
+ " valandtest_orig.append(a1)\n",
+ "\n",
+ "for b1 in test_extract:\n",
+ " valandtest_orig.append(b1)\n",
+ " \n",
+ "for a2 in f1:\n",
+ " valandtest_pred.append(a2)\n",
+ "\n",
+ "for b2 in f2:\n",
+ " valandtest_pred.append(b2)\n"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 27,
+ "id": "lyric-serbia",
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "Text(0.5, 1.0, 'Residuals')"
+ ]
+ },
+ "execution_count": 27,
+ "metadata": {},
+ "output_type": "execute_result"
+ },
+ {
+ "data": {
+ "image/png": "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\n",
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {
+ "needs_background": "light"
+ },
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "residuals = abs(np.array(valandtest_pred)-np.array(valandtest_orig))\n",
+ "residuals_df=pd.DataFrame()\n",
+ "residuals_df['residuals']=residuals\n",
+ "#residuals_df['ds']=prophet_dataset2['ds']\n",
+ "\n",
+ "f, ax = plt.subplots(figsize=(15,7))\n",
+ "ax.text(x=len(val_extract)/2 -10,y=400, s='Validation Data', color='black', fontsize=15, fontweight='bold', alpha=0.8)\n",
+ "ax.text(x=len(val_extract)+10,y=400, s='Testing Data', color='black', fontsize=15, fontweight='bold', alpha=0.8)\n",
+ "ax.axvline(x=len(val_extract), c='green', lw=3, alpha=0.8, ls='--')\n",
+ "residuals_df.plot(kind='line', y='residuals', color='blue', label='Residual',figsize=(15, 6),ax=ax)\n",
+ "plt.title(\"Residuals\")"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "id": "suitable-finding",
+ "metadata": {},
+ "source": [
+ "# Save the predictions"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 28,
+ "id": "herbal-undergraduate",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "predictions_df=pd.DataFrame()\n",
+ "predictions_df['original'] = valandtest_orig\n",
+ "predictions_df['predictions'] = valandtest_pred\n",
+ "predictions_df.to_csv(Path(\"Predictions/CubeRoot.csv\"),index=False)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "id": "voluntary-clearing",
+ "metadata": {},
+ "outputs": [],
+ "source": []
+ }
+ ],
+ "metadata": {
+ "kernelspec": {
+ "display_name": "Python 3",
+ "language": "python",
+ "name": "python3"
+ },
+ "language_info": {
+ "codemirror_mode": {
+ "name": "ipython",
+ "version": 3
+ },
+ "file_extension": ".py",
+ "mimetype": "text/x-python",
+ "name": "python",
+ "nbconvert_exporter": "python",
+ "pygments_lexer": "ipython3",
+ "version": "3.7.10"
+ },
+ "toc-autonumbering": true
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}
diff --git a/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/log.ipynb b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/log.ipynb
new file mode 100644
index 0000000000000000000000000000000000000000..cbf16277e1aa63fc81eba6ec9c5b61923a130a65
--- /dev/null
+++ b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/log.ipynb
@@ -0,0 +1,1955 @@
+{
+ "cells": [
+ {
+ "cell_type": "code",
+ "execution_count": 1,
+ "id": "stuck-potential",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "%matplotlib inline\n",
+ "import mxnet as mx\n",
+ "from mxnet import gluon\n",
+ "import numpy as np\n",
+ "import pandas as pd\n",
+ "import matplotlib.pyplot as plt\n",
+ "import json\n",
+ "from gluonts.dataset.common import ListDataset\n",
+ "from datetime import datetime, timedelta\n",
+ "from gluonts.model.simple_feedforward import SimpleFeedForwardEstimator\n",
+ "from gluonts.model.deepar import DeepAREstimator\n",
+ "from gluonts.mx.trainer import Trainer\n",
+ "from gluonts.evaluation.backtest import make_evaluation_predictions\n",
+ "from gluonts.evaluation import Evaluator\n",
+ "from gluonts.dataset.util import to_pandas\n",
+ "from gluonts.dataset.field_names import FieldName\n",
+ "from gluonts.model.forecast import SampleForecast\n",
+ "from pandas import Timestamp\n",
+ "import itertools\n",
+ "from gluonts.model.estimator import GluonEstimator"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "id": "exclusive-principal",
+ "metadata": {},
+ "source": [
+ "# Data Loading"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 2,
+ "id": "infectious-flour",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "filename='EstimatedRemainingTimeContext'"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 3,
+ "id": "velvet-socket",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "dataset= pd.read_csv(filename + \".csv\")\n",
+ "X= dataset[['name','time', 'countryCode', 'ipAddress', 'level',\n",
+ " 'producer']]\n",
+ "y= dataset.iloc[:,6]\n",
+ "#changing the names of the attributes\n",
+ "gluon_dataset= pd.DataFrame()\n",
+ "gluon_dataset['ds'] = pd.to_datetime(X[\"time\"])\n",
+ "gluon_dataset['y']=y\n",
+ "gluon_dataset['ds'] = gluon_dataset['ds'].dt.tz_convert(None)\n",
+ "gluon_dataset_copy=gluon_dataset.copy()"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 4,
+ "id": "transparent-ethnic",
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/html": [
+ "\n",
+ "\n",
+ "
\n",
+ " \n",
+ " \n",
+ " | \n",
+ " ds | \n",
+ " y | \n",
+ "
\n",
+ " \n",
+ " \n",
+ " \n",
+ " | 0 | \n",
+ " 2021-02-18 11:31:12.752 | \n",
+ " 4050.000000 | \n",
+ "
\n",
+ " \n",
+ " | 1 | \n",
+ " 2021-02-18 11:31:32.752 | \n",
+ " 4045.500000 | \n",
+ "
\n",
+ " \n",
+ " | 2 | \n",
+ " 2021-02-18 11:32:02.752 | \n",
+ " 4045.500000 | \n",
+ "
\n",
+ " \n",
+ " | 3 | \n",
+ " 2021-02-18 11:32:32.752 | \n",
+ " 4045.500000 | \n",
+ "
\n",
+ " \n",
+ " | 4 | \n",
+ " 2021-02-18 11:33:02.752 | \n",
+ " 4153.250000 | \n",
+ "
\n",
+ " \n",
+ " | ... | \n",
+ " ... | \n",
+ " ... | \n",
+ "
\n",
+ " \n",
+ " | 135 | \n",
+ " 2021-02-18 12:38:32.752 | \n",
+ " 110.500000 | \n",
+ "
\n",
+ " \n",
+ " | 136 | \n",
+ " 2021-02-18 12:39:02.752 | \n",
+ " 60.156250 | \n",
+ "
\n",
+ " \n",
+ " | 137 | \n",
+ " 2021-02-18 12:39:32.752 | \n",
+ " 19.140625 | \n",
+ "
\n",
+ " \n",
+ " | 138 | \n",
+ " 2021-02-18 12:40:02.752 | \n",
+ " 5.359375 | \n",
+ "
\n",
+ " \n",
+ " | 139 | \n",
+ " 2021-02-18 12:40:32.752 | \n",
+ " 0.765625 | \n",
+ "
\n",
+ " \n",
+ "
\n",
+ "
140 rows × 2 columns
\n",
+ "
"
+ ]
+ },
+ "metadata": {
+ "needs_background": "light"
+ },
+ "output_type": "display_data"
+ },
+ {
+ "data": {
+ "image/png": "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\n",
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {
+ "needs_background": "light"
+ },
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "gluon_dataset_copy['y_orig']=gluon_dataset_copy['y']\n",
+ "transformed = log_transformation(gluon_dataset_copy['y'])\n",
+ "gluon_dataset_copy['y']=transformed\n",
+ "\n",
+ "#plot before transformation\n",
+ "pd.plotting.register_matplotlib_converters()\n",
+ "f, ax = plt.subplots(figsize=(15,7))\n",
+ "gluon_dataset_copy.plot(kind='line', x='ds', y='y_orig', color='blue', label='Train', ax=ax)\n",
+ "plt.title(filename)\n",
+ "plt.show()\n",
+ "\n",
+ "#plot after transformation\n",
+ "pd.plotting.register_matplotlib_converters()\n",
+ "f, ax = plt.subplots(figsize=(15,7))\n",
+ "gluon_dataset_copy.plot(kind='line', x='ds', y='y', color='blue', label='Train', ax=ax)\n",
+ "#test.plot(kind='line', x='ds', y='y', color='red', label='Test', ax=ax)\n",
+ "plt.title(filename + \"_transformed using Square Root Transformation\")\n",
+ "plt.show()"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 8,
+ "id": "brazilian-bailey",
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/html": [
+ "\n",
+ "\n",
+ "
\n",
+ " \n",
+ " \n",
+ " | \n",
+ " ds | \n",
+ " y | \n",
+ " y_orig | \n",
+ "
\n",
+ " \n",
+ " \n",
+ " \n",
+ " | 0 | \n",
+ " 2021-02-18 11:31:12.752 | \n",
+ " 8.306719 | \n",
+ " 4050.000000 | \n",
+ "
\n",
+ " \n",
+ " | 1 | \n",
+ " 2021-02-18 11:32:12.752 | \n",
+ " 8.305608 | \n",
+ " 4045.500000 | \n",
+ "
\n",
+ " \n",
+ " | 2 | \n",
+ " 2021-02-18 11:33:12.752 | \n",
+ " 8.305608 | \n",
+ " 4045.500000 | \n",
+ "
\n",
+ " \n",
+ " | 3 | \n",
+ " 2021-02-18 11:34:12.752 | \n",
+ " 8.305608 | \n",
+ " 4045.500000 | \n",
+ "
\n",
+ " \n",
+ " | 4 | \n",
+ " 2021-02-18 11:35:12.752 | \n",
+ " 8.331887 | \n",
+ " 4153.250000 | \n",
+ "
\n",
+ " \n",
+ " | ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ "
\n",
+ " \n",
+ " | 135 | \n",
+ " 2021-02-18 13:46:12.752 | \n",
+ " 4.714025 | \n",
+ " 110.500000 | \n",
+ "
\n",
+ " \n",
+ " | 136 | \n",
+ " 2021-02-18 13:47:12.752 | \n",
+ " 4.113432 | \n",
+ " 60.156250 | \n",
+ "
\n",
+ " \n",
+ " | 137 | \n",
+ " 2021-02-18 13:48:12.752 | \n",
+ " 3.002739 | \n",
+ " 19.140625 | \n",
+ "
\n",
+ " \n",
+ " | 138 | \n",
+ " 2021-02-18 13:49:12.752 | \n",
+ " 1.849930 | \n",
+ " 5.359375 | \n",
+ "
\n",
+ " \n",
+ " | 139 | \n",
+ " 2021-02-18 13:50:12.752 | \n",
+ " 0.568505 | \n",
+ " 0.765625 | \n",
+ "
\n",
+ " \n",
+ "
\n",
+ "
140 rows × 3 columns
\n",
+ "
"
+ ]
+ },
+ "metadata": {
+ "needs_background": "light"
+ },
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "train_entry = next(iter(train_ds))\n",
+ "train_entry.keys()\n",
+ "\n",
+ "validation_entry = next(iter(validation_ds))\n",
+ "validation_entry.keys()\n",
+ "\n",
+ "test_entry = next(iter(test_ds))\n",
+ "test_entry.keys()\n",
+ "\n",
+ "test_orig_entry = next(iter(test_orig_ds))\n",
+ "test_orig_entry.keys()\n",
+ "\n",
+ "test_series = to_pandas(test_entry)\n",
+ "train_series = to_pandas(train_entry)\n",
+ "validation_series = to_pandas(validation_entry)\n",
+ "\n",
+ "prediction_length = len(test_series) - len(validation_series)\n",
+ "context_length = [50,60]\n",
+ "\n",
+ "\n",
+ "list1=train_series.index\n",
+ "list2=train_entry['target']\n",
+ "\n",
+ "d = {'ds':list1,'y':list2}\n",
+ "df_train=pd.DataFrame(d, columns=['ds','y'])\n",
+ "\n",
+ "fig, ax = plt.subplots( figsize=(10, 7))\n",
+ "\n",
+ "# train_series.plot(ax=ax[0])\n",
+ "# ax[0].grid(which=\"both\")\n",
+ "# ax[0].legend([\"train series\"], loc=\"upper left\")\n",
+ "\n",
+ "test_series.plot(ax=ax)\n",
+ "plt.axvline(train_series.index[-1], color='r') # end of train dataset\n",
+ "plt.axvline(validation_series.index[-1], color='b') # end of validation dataset\n",
+ "ax.grid(which=\"both\")\n",
+ "ax.legend([\"test series\", \"end of train series\",\"end of validation series\"], loc=\"upper left\")\n",
+ "\n",
+ "plt.show()"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 12,
+ "id": "interracial-saying",
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "21"
+ ]
+ },
+ "execution_count": 12,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "prediction_length"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "id": "demographic-procedure",
+ "metadata": {},
+ "source": [
+ "# Hyper-parameter Tuning"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 13,
+ "id": "collective-intervention",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "batch_size = [75,100]\n",
+ "epochs = [5,10]\n",
+ "num_batches_per_epoch = [10,20]\n",
+ "learning_rate = [1e-3,1e-4]"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 14,
+ "id": "chronic-matthew",
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "Total Possible Models 32\n"
+ ]
+ }
+ ],
+ "source": [
+ "from sklearn.model_selection import ParameterGrid\n",
+ "\n",
+ "param_grid = {\n",
+ " 'batch_size': batch_size,\n",
+ " 'epochs': epochs,\n",
+ " 'num_batches_per_epoch': num_batches_per_epoch,\n",
+ " 'learning_rate': learning_rate,\n",
+ " 'context_length': context_length\n",
+ " }\n",
+ "grid = ParameterGrid(param_grid)\n",
+ "cnt = 0\n",
+ "for p in grid:\n",
+ " cnt = cnt+1\n",
+ "\n",
+ "print('Total Possible Models',cnt)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 15,
+ "id": "paperback-thesis",
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stderr",
+ "output_type": "stream",
+ "text": [
+ " 0%| | 0/10 [00:00"
+ ]
+ },
+ "metadata": {
+ "needs_background": "light"
+ },
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "def plot_prob_forecasts(ts_entry, forecast_entry):\n",
+ " plot_length = len(ts_entry)\n",
+ " prediction_intervals = (50.0,98.0)\n",
+ " legend = [\"observations\", \"median prediction\"] + [f\"{k}% prediction interval\" for k in prediction_intervals][::-1]\n",
+ "\n",
+ " fig, ax = plt.subplots(1, 1, figsize=(10, 7))\n",
+ " ts_entry[-plot_length:].plot(ax=ax) # plot the time series\n",
+ " forecast_entry.plot(prediction_intervals=prediction_intervals, color='g')\n",
+ " plt.grid(which=\"both\")\n",
+ " plt.legend(legend, loc=\"upper left\")\n",
+ " plt.show()\n",
+ "plot_prob_forecasts(gluon_dataset['y'], forecast_entry_inv)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 22,
+ "id": "experimental-jonathan",
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stderr",
+ "output_type": "stream",
+ "text": [
+ "Running evaluation: 100%|██████████| 1/1 [00:00<00:00, 5.89it/s]\n"
+ ]
+ }
+ ],
+ "source": [
+ "evaluator = Evaluator(quantiles=[0.1, 0.5, 0.9])\n",
+ "agg_metrics, item_metrics = evaluator(iter(tss_inv), iter([forecast_entry_inv]), num_series=len(test_ds))\n"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 23,
+ "id": "otherwise-winner",
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "{'MSE': 28775.041666666668,\n",
+ " 'abs_error': 2826.3955078125,\n",
+ " 'abs_target_sum': 4846.515625,\n",
+ " 'abs_target_mean': 230.78645833333334,\n",
+ " 'seasonal_error': 86.95240499205508,\n",
+ " 'MASE': 1.547861295947862,\n",
+ " 'MAPE': 23.938071251136698,\n",
+ " 'sMAPE': 0.6052190021513292,\n",
+ " 'OWA': nan,\n",
+ " 'MSIS': 25.965856988129495,\n",
+ " 'QuantileLoss[0.1]': 2810.276113891602,\n",
+ " 'Coverage[0.1]': 0.6190476190476191,\n",
+ " 'QuantileLoss[0.5]': 2826.395523071289,\n",
+ " 'Coverage[0.5]': 1.0,\n",
+ " 'QuantileLoss[0.9]': 925.3613464355468,\n",
+ " 'Coverage[0.9]': 1.0,\n",
+ " 'RMSE': 169.6320773517399,\n",
+ " 'NRMSE': 0.735017464095463,\n",
+ " 'ND': 0.583180933789417,\n",
+ " 'wQuantileLoss[0.1]': 0.5798549579403455,\n",
+ " 'wQuantileLoss[0.5]': 0.583180936937821,\n",
+ " 'wQuantileLoss[0.9]': 0.1909333257200727,\n",
+ " 'mean_absolute_QuantileLoss': 2187.3443277994793,\n",
+ " 'mean_wQuantileLoss': 0.4513230735327464,\n",
+ " 'MAE_Coverage': 0.373015873015873}"
+ ]
+ },
+ "execution_count": 23,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "agg_metrics"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "id": "premier-healthcare",
+ "metadata": {},
+ "source": [
+ "# Residual"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 24,
+ "id": "hydraulic-religion",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "#get the mean for all the samples\n",
+ "f1 = forecast_entry_val_inv.samples.mean(axis=0)\n",
+ "f2 = forecast_entry_inv.samples.mean(axis=0)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 25,
+ "id": "assisted-fault",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "original_values=list(gluon_dataset['y'])\n",
+ "orig_test = original_values[:len(original_values)-prediction_length]\n",
+ "test_extract=original_values[-prediction_length:]\n",
+ "val_extract=orig_test[-prediction_length:]"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 26,
+ "id": "assured-siemens",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "valandtest_orig=list()\n",
+ "valandtest_pred=list()\n",
+ "for a1 in val_extract:\n",
+ " valandtest_orig.append(a1)\n",
+ "\n",
+ "for b1 in test_extract:\n",
+ " valandtest_orig.append(b1)\n",
+ " \n",
+ "for a2 in f1:\n",
+ " valandtest_pred.append(a2)\n",
+ "\n",
+ "for b2 in f2:\n",
+ " valandtest_pred.append(b2)\n"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 29,
+ "id": "decimal-viewer",
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "Text(0.5, 1.0, 'Residuals')"
+ ]
+ },
+ "execution_count": 29,
+ "metadata": {},
+ "output_type": "execute_result"
+ },
+ {
+ "data": {
+ "image/png": "iVBORw0KGgoAAAANSUhEUgAAA20AAAF1CAYAAAB76AIVAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuNCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8QVMy6AAAACXBIWXMAAAsTAAALEwEAmpwYAABmNklEQVR4nO3dd3hU1dbH8e+mCQhIFSkqoCjYEIkgikgRGzYQFTtXNKigeL167Q3Lq1cRxYbYAMGKFAUVlIhINyAoCkREkN6kd5L9/rEyIUCASTKTMzP5fZ5nnplMOWdNMknOOnvttZ33HhEREREREYlNRYIOQERERERERPZPSZuIiIiIiEgMU9ImIiIiIiISw5S0iYiIiIiIxDAlbSIiIiIiIjFMSZuIiIiIiEgMU9ImIiKFgnPuOufc6AM8PtY5d0sE9tPCObc4v9sREREJUdImIiIxyTm3wDm31Tm3yTm33DnXzzlXJq/b894P8t6fF8kYRURECoKSNhERiWWXeO/LAKcCDYEHgw1HRESk4ClpExGRmOe9Xw6MwpI3nHNnOOcmOufWOedmOudahJ7rnOvknJvvnNvonPvLOXddtvvHZ3teG+fcHOfceufca4DL9tgTzrmB2b6u5ZzzzrlimV//yzk3O3Mf851zXfYXu3PufufcksznznXOtY7U90VERAoHJW0iIhLznHM1gQuBec65GsBI4GmgInAv8Llzropz7lCgN3Ch974scCYwI4ftVQaGAI8AlYE/gbNyEdJK4GKgHPAvoJdz7rQc9nM80A04PTOe84EFudiPiIiIkjYREYlpw5xzG4FFWKL0OHA98JX3/ivvfYb3/lsgFbgo8zUZwEnOuVLe+2Xe+99y2O5FwG/e+8He+53Ay8DycIPy3o/03v/pzQ/AaODsHJ6aDhwCnOCcK+69X+C9/zPc/YiIiICSNhERiW2XZ45QtQDqYaNiRwNXZpZGrnPOrQOaAdW895uBq4HbgGXOuZHOuXo5bLc6lggC4L332b8+GOfchc65yc65fzL3f1FmbHvw3s8D7gaeAFY65z52zlUPdz8iIiKgpE1EROJA5mhWP+BFLLn6wHtfPtvlUO/9c5nPHeW9bwNUA+YAb+ewyWXAkaEvnHMu+9fAZqB0tq+PyPbcQ4DPM2Op6r0vD3xFtjlxe8X+ofe+GZZseuD5XLx1ERERJW0iIhI3XgbaABOBS5xz5zvnijrnSmaujVbTOVfVOXdZ5ty27cAmrFxybyOBE51z7TObi9xFtsQMmwfX3Dl3lHPuMPbsWlkCK3lcBexyzl0I5LiUgHPueOdcq8xEbxuwdT/xiIiI7JeSNhERiQve+1XAACzBugx4CEucFgH3Yf/TigD3AEuBf4BzgNtz2NZq4ErgOWANUBeYkO3xb4FPgF+AacCIbI9tzIzhU2AtcC3wxX7CPiRzH6uxOXOHo2ULREQkl5yV8YuIiIiIiEgs0kibiIiIiIhIDFPSJiIiIiIiEsOUtImIiIiIiMQwJW0iIiIiIiIxTEmbiIiIiIhIDCsWdAAAlStX9rVq1Qo6DBERkf1atWVV1u0qpasEGImIiCSiadOmrfbe5/gPJiaStlq1apGamhp0GCIiIvuV1Dcp63Zqsv5niYhIZDnnFu7vMZVHioiIiIiIxDAlbSIiIiIiIjFMSZuIiIiIiEgMi4k5bTnZuXMnixcvZtu2bUGHEldKlixJzZo1KV68eNChiIiIiIhIBMRs0rZ48WLKli1LrVq1cM4FHU5c8N6zZs0aFi9eTO3atYMOR0REREREIiBmyyO3bdtGpUqVlLDlgnOOSpUqaXRSRERERCSBxOxIG6CELQ/0PRMRiY7kRslBhyAiIoVUTCdtQStatCgnn3wyu3btonbt2nzwwQeUL18+V9tITU1lwIAB9O7de5/HQuvTVa5cOdexPfHEE5QpU4Z77703168VEZHcU9ImIiJBidnyyFhQqlQpZsyYwaxZs6hYsSKvv/56rreRlJSUY8ImIiIiIiISDiVtYWratClLliwB4M8//+SCCy6gUaNGnH322cyZMweAzz77jJNOOokGDRrQvHlzAMaOHcvFF18MwJo1azjvvPM48cQTueWWW/DeA7BgwQJOOumkrH29+OKLPPHEEwC8/fbbnH766TRo0IArrriCLVu2FNRbFhERERGRGBAX5ZF33w0zZkR2m6eeCi+/HN5z09PTGTNmDJ07dwYgOTmZPn36ULduXaZMmcIdd9xBSkoKPXr0YNSoUdSoUYN169bts50nn3ySZs2a8dhjjzFy5Ejefffdg+67ffv23HrrrQA88sgjvPvuu9x5551hvksREREREYl3cZG0BWXr1q2ceuqpLFmyhPr169OmTRs2bdrExIkTufLKK7Oet337dgDOOussOnXqxFVXXUX79u332d64ceMYMmQIAG3btqVChQoHjWHWrFk88sgjrFu3jk2bNnH++edH6N2JiEhu/Pubf2fd7nVBrwAjERGRvPAe0tJgwQKIt0PquEjawh0Ri7TQnLYtW7Zw/vnn8/rrr9OpUyfKly/PjByG/vr06cOUKVMYOXIkjRo1Ytq0aWHtp1ixYmRkZGR9nb1lf6dOnRg2bBgNGjSgX79+jB07Nr9vS0RE8uDHv38MOgQREcmF9HT49VcYN84uP/4IK1dChQqwejUUiaOJYnEUanBKly5N79696dmzJ6VLl6Z27dp89tlngC1oPXPmTMDmujVp0oQePXpQpUoVFi1atMd2mjdvzocffgjA119/zdq1awGoWrUqK1euZM2aNWzfvp0RI0ZkvWbjxo1Uq1aNnTt3MmjQoIJ4uyIiIiIicWfHDpg0CZ5/Hi6+GCpVgoYNoXt3SE2F886Dvn3tOfG2SlZcjLTFgoYNG3LKKafw0UcfMWjQIG6//Xaefvppdu7cSceOHWnQoAH33Xcff/zxB957WrduTYMGDfjhhx+ytvH4449zzTXXcOKJJ3LmmWdy1FFHAVC8eHEee+wxGjduTI0aNahXr17Wa5566imaNGlClSpVaNKkCRs3bizw9y4iIiIiEmu2bIHJk20Ebdw4S8a2brXH6tWDq66C5s3h7LPh6KODjTW/XKiDYZCSkpJ8amrqHvfNnj2b+vXrBxRRfNP3TkQk8pL6JmXdTk1OPcAzRUQkGtavhwkTdpc7pqbCzp02anbqqZacNW8OzZpB1apBR5t7zrlp3vuknB7TSJuIiIiIiMSkjAwYPBheeAGmTbNmIsWLQ1IS3HOPJWlnngnlywcdaXQpaRMRERERkZjiPYwcCY88AjNnwgknwOOPW5LWpAmULh10hAVLSZuIiIiIiMSMlBR4+GGbr1anDgwcCB07QtGiQUcWnJjuHhkL8+3ijb5nIiIiIhKPJk+G1q3tsngxvPUWzJkD111XuBM2iOGkrWTJkqxZs0ZJSC5471mzZg0lS5YMOhQRERERkbDMmAGXXAJNm9q6ar16wR9/QHKyzV+TGC6PrFmzJosXL2bVqlVBhxJXSpYsSc2aNYMOQ0RERETkgObMsXlqn35qjUSefRbuvBPKlAk6stgTs0lb8eLFqV27dtBhiIiIAPDQ2Q8FHYKISEJYsACefBIGDIBSpWz+2r33Jn4HyPyI2aRNREQklrSv3z7oEERE4trSpfDMM/D221CkCHTvDg88AIcfHnRksU9Jm4iIiIiIRM3q1fD88/Daa7BrF3TubK38NaMnfEraREREREQk4jZsgJ49rbHIpk1w/fXwxBPWxl9yR0mbiIiIiIhE3FVXwahRcMUV0KOHLZAteaOkTUREJAzXD7k+6/bA9gMDjEREJPbNmWMJW48e8OijQUcT/5S0iYiIhGHO6jlBhyAiEjfeesvWWEtODjqSxBCzi2uLiIiIiEj82boV+vWD9u2hatWgo0kMStpERERERCRiPvkE1q2D224LOpLEoaRNREREREQipk8fqFcPzjkn6EgSx0GTNudcSefcVOfcTOfcb865JzPvr+2cm+Kcm+ec+8Q5VyLz/kMyv56X+XitKL8HERERERGJAT//DFOm2Cibc0FHkzjCGWnbDrTy3jcATgUucM6dATwP9PLeHwusBTpnPr8zsDbz/l6ZzxMRERERkQTXpw+UKgU33hh0JInloEmbN5syvyyeefFAK2Bw5v39gcszb1+W+TWZj7d2Tnm2iIiIiEgi27ABBg2Cjh2hQoWgo0ksYc1pc84Vdc7NAFYC3wJ/Auu897syn7IYqJF5uwawCCDz8fVApQjGLCIiIiIiMWbgQNi8GW6/PehIEk9YSZv3Pt17fypQE2gM1Mvvjp1zyc65VOdc6qpVq/K7ORERERERCYj38OabcNppkJQUdDSJJ1fdI73364DvgaZAeedcaHHumsCSzNtLgCMBMh8/DFiTw7b6eu+TvPdJVapUyVv0IiIiIiISuIkTYdYsG2XTxKjIK3awJzjnqgA7vffrnHOlgDZYc5HvgQ7Ax8BNwPDMl3yR+fWkzMdTvPc+CrGLiIgUmJfOfynoEEREYtabb0K5cnDNNUFHkpgOmrQB1YD+zrmi2Mjcp977Ec6534GPnXNPAz8D72Y+/13gA+fcPOAfoGMU4hYRESlQzY9uHnQIIiIxafVq+OwzSE6GQw8NOprEdNCkzXv/C9Awh/vnY/Pb9r5/G3BlRKITEREREZGY9v77sGOHrc0m0ZGrOW0iIiIiIiIhGRnw1ltw9tlw4olBR5O4lLSJiIiIiEiefPcd/Pmn2vxHWzhz2kRERAq9CwZekHX7m+u/CTASEZHY8eabUKUKtG8fdCSJTUmbiIhIGFZvWR10CCIiMWXxYvjyS7j3XjjkkKCjSWwqjxQRERERkVx75x2b05acHHQkiU9Jm4iIiIiI5MquXfD223D++VCnTtDRJD4lbSIiIiIikitffglLl6rNf0FR0iYiIiIiIrnSpw/UrAlt2wYdSeGgpE1ERERERMI2bx6MHm1z2YqprWGBUNImIiIiIiJh69sXihaFzp2DjqTwUNImIiIiIiJh2bYN3nsPLr8cqlcPOprCQ0mbiIiIiIiE5fPPYc0aNSApaEraREREREQkLG++CXXrQqtWQUdSuGjqoIiISBg+aPdB0CGIiATq119hwgR48UUooqGfAqWkTUREJAz1q9QPOgQRkUD16QOHHAKdOgUdSeGjHFlERERERA5o0yb44AO46iqoVCnoaAofJW0iIiIiInJAH34IGzfC7bcHHUnhpKRNRERERET2y3trQNKgAZxxRtDRFE6a0yYiIhKGpL5JWbdTk1MDjEREpGBNnQozZlji5lzQ0RROGmkTEREREZH9evNNKFMGrrsu6EgKLyVtIiIiIiKSo3/+gU8+geuvh7Jlg46m8FLSJiIiIiIiOerfH7Ztg9tuCzqSwk1Jm4iIiIiI7MN7W5utaVNrQiLBUSMSERERERHZx/ffQ1oaDBgQdCSikTYREREREdlHnz5QsSJceWXQkYiSNhERERER2cOyZTB0KPzrX1CyZNDRiJI2ERERERHZw3vvwa5d0KVL0JEIKGkTEREREZFs0tOhb18491yoWzfoaASUtImIiIiISDZffw1//w233x50JBKi7pEiIiJh+Pq6r4MOQUSkQLz5JlSrBpdcEnQkEqKkTUREJAxVDq0SdAgiIlG3YIGNtD36KBQvHnQ0EqLySBERERERAaB/f7vu3DnYOGRPStpERERERISMDEvazj0Xjjoq6GgkOyVtIiIiYVi1eVXWRUQkEf34I/z1F3TqFHQksjfNaRMREQnDhYMuzLqdmpwaYCQiItHRrx+UKweXXx50JLI3jbSJiIiIiBRymzbBZ5/B1VdD6dJBRyN7U9ImIiIiIlLIff45bN6s0shYddCkzTl3pHPue+fc786535xz3TPvf8I5t8Q5NyPzclG21zzonJvnnJvrnDs/mm9ARERERETyp18/qFsXmjYNOhLJSThz2nYB//HeT3fOlQWmOee+zXysl/f+xexPds6dAHQETgSqA985547z3qdHMnAREREREcm/v/6CsWPh6afBuaCjkZwcdKTNe7/Mez898/ZGYDZQ4wAvuQz42Hu/3Xv/FzAPaByJYEVEREREJLIGDLBk7YYbgo5E9idXc9qcc7WAhsCUzLu6Oed+cc6955yrkHlfDWBRtpct5sBJnoiIiIiIBCC0Nlvr1lqbLZaFnbQ558oAnwN3e+83AG8CxwCnAsuAnrnZsXMu2TmX6pxLXbVKa96IiIiIiBQ0rc0WH8JK2pxzxbGEbZD3fgiA936F9z7de58BvM3uEsglwJHZXl4z8749eO/7eu+TvPdJVapUyc97EBERERGRPOjXD8qWhXbtgo5EDiSc7pEOeBeY7b1/Kdv91bI9rR0wK/P2F0BH59whzrnaQF1gauRCFhERERGR/NLabPEjnO6RZwE3AL8652Zk3vcQcI1z7lTAAwuALgDe+9+cc58Cv2OdJ7uqc6SIiIiISGzR2mzxw3nvg46BpKQkn5qaGnQYIiIiIiKFRsuWsHgxpKWp1X8scM5N894n5fRYrrpHioiIiIhI/AutzdapkxK2eKCkTURERESkkNHabPFFSZuIiIiISCGitdniTziNSERERAq92atmZ92uX6V+gJGIiORPaG22p54KOhIJl5I2ERGRMNwwdHcNUWqymmeJSPzS2mzxR+WRIiIiIiKFhNZmi09K2kRERERECgmtzRaflLSJiIiIiBQS/fvDscfCmWcGHYnkhpI2EREREZFCYMEC+P57rc0Wj5S0iYiIiIgUAlqbLX4paRMRERERSXAZGdY1slUrrc0Wj5S0iYiIiIgkuPHjbW02NSCJT0raREREREQSnNZmi29K2kREREREEtimTfDpp3DVVXDooUFHI3lRLOgARERE4kHl0pWDDkFEJE+GDNHabPFOSZuIiEgYvrn+m6BDEBHJk3794Jhj4Kyzgo5E8krlkSIiIiIiCUprsyUGJW0iIiIiIgkqtDbbjTcGHYnkh5I2EREREZEEpLXZEofmtImIiIRh3MJxWbebH908wEhERMITWputR4+gI5H8UtImIiIShntG3ZN1OzU5NcBIRETCo7XZEofKI0VEREREEozWZkssStpERERERBKM1mZLLEraREREREQSjNZmSyxK2kREREREEojWZks8StpERERERBKI1mZLPEraREREREQShNZmS0xK2kREREREEkRobTY1IEksStpERERERBJEv35QpozWZks0StpERERERBKA1mZLXMWCDkBERCQe1KtcL+gQREQOSGuzJS4lbSIiImEY2H5g0CGIiBxQv35Qpw40axZ0JBJpKo8UEREREYlzCxdqbbZEpqRNRERERCTODRhg11qbLTEpaRMRERERiWNjxkDPntC6NRx9dNDRSDRoTpuIiEgYhsweknW7ff32AUYiIrJbnz7QrRvUrw/vvBN0NBItStpERETC8OyPz2bdVtImIkHbtQv+/W947TVo2xY+/BDKlQs6KomWg5ZHOueOdM5975z73Tn3m3Oue+b9FZ1z3zrn/si8rpB5v3PO9XbOzXPO/eKcOy3ab0JEREREpLBYt84Stddeg//8B4YPV8KW6MKZ07YL+I/3/gTgDKCrc+4E4AFgjPe+LjAm82uAC4G6mZdk4M2IRy0iIiIiUgjNmwdnnGGdIt99F158EYoWDToqibaDJm3e+2Xe++mZtzcCs4EawGVA/8yn9Qcuz7x9GTDAm8lAeedctUgHLiIiIiJSmHz/PTRuDKtXw3ffwc03Bx2RFJRcdY90ztUCGgJTgKre+2WZDy0HqmbergEsyvayxZn3iYiIiIhIHvTtC+edB9WqwdSp0Lx50BFJQQo7aXPOlQE+B+723m/I/pj33gM+Nzt2ziU751Kdc6mrVq3KzUtFRERERAqFXbvg7ruhSxdo0wYmToQ6dYKOSgpaWEmbc644lrAN8t6Heh6vCJU9Zl6vzLx/CXBktpfXzLxvD977vt77JO99UpUqVfIav4iIiIhIQlq/Hi6+GF55xRK3L7+Eww4LOioJQjjdIx3wLjDbe/9Stoe+AG7KvH0TMDzb/TdmdpE8A1ifrYxSREREREQO4s8/oWlTWzi7b1/o1UsNRwqzcNZpOwu4AfjVOTcj876HgOeAT51znYGFwFWZj30FXATMA7YA/4pkwCIiIiIiieyHH6B95nKQ334LLVoEGo7EgIMmbd778YDbz8Otc3i+B7rmMy4RERERkULnnXfg9tvh2GOtHPLYY4OOSGJBOCNtIiIihd7ZR50ddAgiksDS0+G++6wM8rzz4JNPoHz5oKOSWKGkTUREJAy9LugVdAgikqA2bIBrroGvvoI774SXXoJiOkqXbPRxEBEREREJyPz5cMklMHcuvPkm3HZb0BFJLFLSJiIiIiISgPHjoV07K40cPRpatQo6IolVYS+uLSIiIiIikfHJJ9C6NVSoAJMnK2GTA9NIm4iISBj6TuubdTu5UXKAkYhIPPMenn8eHnwQmjWDYcOgUqWgo5JYp6RNREQkDEraRCS/du6Erl3h7behY0d4/30oWTLoqCQeqDxSRERERCTKNmywhiNvvw0PPQSDBilhk/BppE1EREREJIoWL4a2beG33yxpu+WWoCOSeKOkTUREREQkSmbMsIRt40Zbh+2884KOSOKRyiNFRERERKLg66/h7LOhSBGYMEEJm+SdkjYRERERkQh76y2bw3bssTBlCpx8ctARSTxT0iYiIiIiEiEZGXD//XDbbXD++TBuHFSvHnRUEu80p01EREREJAK2boWbboLPPrOk7dVXoZiOtiUC9DESEREREcmn1avhsstg4kR44QX4z3/AuaCjkkShpE1EREREJB/++AMuusha+3/2GXToEHREkmiUtImIiIShXb12QYcgIjFo/Hi4/HIbVUtJgaZNg45IEpGSNhERkTA83PzhoEMQkRjzySdw441Qq5atwXbMMUFHJIlK3SNFRERERHLBe3j+eejYERo3tnlsStgkmpS0iYiIiIiEyXvo2hUeeACuuQa+/RYqVQo6Kkl0StpERERERML06qvw5ptw770wcCCULBl0RFIYaE6biIhIGJ4Z90zWbc1vEymcJk2yVv6XXmrlkUU0/CEFREmbiIhIGIbOGZp1W0mbSOGzahVcdRUceST066eETQqWkjYRERERkQNIT4frr7fEbeJEqFAh6IiksFHSJiIiIiJyAE89BaNHQ9++cNppQUcjhZEGdkVERERE9mPUKOjRw9Zju+WWoKORwkpJm4iIiIhIDv7+G667Dk46yTpGOhd0RFJYKWkTEREREdnLjh3WeGTHDhg8GEqXDjoiKcw0p01EREREZC/33gtTpsBnn8FxxwUdjRR2GmkTEREREcnmk09sEe2774YOHYKORkQjbRGRnJzM9OnTKVKkCKNHj6Z8+fJ7PD5gwAB69+4NQI8ePbjooovC2m5SUhIAF198MU888QQAl1xyCcuWLeO0006jb9+++33ttGnT6NKlCwCPP/44l1xySa7e04cffsimTZuoVq3aPq8NN4Zo6Nu3b9Y+nXMUL16csmXLcvTRR3POOefQrl07SuexfiG03eOOO44WLVpEKmQREZG4Evo/fzBvvfUWjRo1isg+Y+m4Y84cazhy5plw991LSUq6NOuxIkWKULp0aapWrcqpp55Kx44dqV27dp72M3bsWNLS0gA7lhQ5EI20RcC5554LQEZGBt9///0+j3/33XcAlChRIm6SgY8++oi+ffvy5ZdfBh3Kfnnv2bFjB2vWrGH69On06tWLG2+8keXLl+dpe6GEcOzYsZENVERERA4oVo47Nm+2kbWSJW20rXjxPR/PyMhg06ZN/Pnnn3z++edce+21fP3113na19ixY/c4GS1yIBppi4Bzzz2XF198kYyMDMaMGUO7du2yHlu+fDm///47AE2bNs3zKFBI0H/MYiUGsBHEiy66iEWLFvHee+/x1VdfsWDBAu69914GDBhAkSI6JyEikZPcSGfCpXDI/n8+e+VO9sqfoOKJJu+hSxf4/Xdbk61mTVi6dPfjoZG+devWkZKSQq9evdi6dStPPvkkRx99NCeccEKBxCmFk5K2CKhYsSKnnXYaqamppKamsmHDBsqVKwdASkpK1vPOO+88AHr16sVPP/3E8uXL2bx5M2XLluWUU07htttu47iDzHTNqUQgIyOD119/neHDh7Njxw5atmxJq1at9nlteno6Tz31FL///jurVq1iy5YtlC9fnqSkJLp27Ur16tVZunQpl166uwxg+vTpWWWaycnJJCcn77dMYdy4cQwcOJA5c+awa9cujjzySC6++GKuu+66rAQq+xmlvn37MnDgQKZOnUr58uW56qqruPHGG3P1vS9atCi1atWiR48eLFmyhJkzZzJnzhzGjRtHixYtWLt2LS+88AJz5sxh9erV7Nixg0qVKnHWWWdxxx13UL58+T3+IQGMGDGCESNGALtLS/PzMxORxKCkTWRPK1eu5O2332bChAn8888/lC9fPuv/a6VKlbKe99lnnzF06FAWL15MRkYGlSpVon79+iQnJ1OyZMk8HXeEpqZUq1aNJ554gldeeYV58+Zx5JFH8u9//5szzjgja5tLlizh2Wef5eeff6ZKlSp07tyZ6dOnZ/2vT01NBeCtt2DQIFuTLbOIKkfly5enffv2FClShKeffppdu3bx7rvv0rNnTwD69+9PSkoKS5YsYePGjZQuXZr69etz8803Z723vUtQQ/eH3uO4ceP4+OOPmT9/PuvXr8863mnfvj3t27fP889M4peGIiKkTZs2AOzatWuP8roxY8YAcMghh3D22WcD8PXXX5OWlsaGDRtIT09n3bp1jBs3juTkZFavXp3rffft25f+/fuzbt06tmzZwsiRI3n22Wf3eV56ejojRoxg/vz5bNy4kfT0dNasWcOoUaPo0qULO3bsyMM7N4MHD+aee+5h+vTpbNmyhR07dvDnn3/yyiuv8PDDD+f4mnvuuYdx48axbds2li9fTu/evZk4cWKeY+jYsWPW7dB21q9fz+jRo/n777/ZsmULu3btYsWKFQwZMoTu3buHve1I/8xERETi2cqVK7nxxhsZOnQoK1euZNeuXaxevZrhw4fTqVMn1q1bB8Do0aN5/vnnSUtLY8uWLWzbto0lS5bw3Xff8ddff+U7jn/++Ydu3brx+++/Zx173Hfffaxfvx6AnTt30rVrV6ZMmcKOHTtYsmQJPXr0YPLkyXtsJzUVuneHCy+E/Ry27OPSSy/NqqCaOnUqGRkZAHz//ff89ttvrFu3jvT0dDZu3MjUqVPp1q1b1hy2g5k5cyZTp05l9erV7Ny5k23btjFnzhyeffZZBg8eHOZ3RxKJkrYIadWqVdZoUihRW716Nb/88gsAZ511VtYv9v3338+wYcMYP348EydO5LXXXgNg06ZNjBo1Klf73bhxIx9++CEARxxxBIMHD2bkyJFUrFhxn+cWK1aMZ599lhEjRjBx4kTGjx/Po48+CsCyZcsYP3481atXJzU1lWrVqgHsMYK4v0myW7ZsyWq0cvjhh/PRRx8xevRoTj/9dAC+/fZbpk2bts/ratWqxddff531/rN/7/Li6KOPzrodmtdWsWJFevbsyddff82kSZP44Ycfst7Hb7/9xpw5c2jUqFHWWTaw8o/Qew5Nho7kz0xERCTe9enTh9WrV1O+fHn69evHpEmTeOeddyhRogTLli3jgw8+AGDGjBkA1KhRg1GjRjFhwgQGDx7MfffdR7Vq1fJ03JHd9u3bueyyy0hJSeGWW24BYOvWrUyYMAGAkSNHsnjxYsBGt77//nv+97//sWbNmqxt/POPzWM74gj44AMId3ZFkSJFqFmzZtY+Q4nqLbfcwuDBgxk3bhyTJ0/mgw8+oGTJkuzatYvhw4cDVvJ58cUXZ20r9J5DI4ktWrRgwIABpKSkMGXKFEaMGEG9evUAlLQVUiqPjJAKFSqQlJTE1KlTmTp1Kps2bSIlJQXvPbB7JA6gePHi9OjRg7S0NDZv3pz1HICFCxfmar9//PEHW7ZsAeDyyy+nVq1aAFxzzTX06NFjj+cWKVKEbdu28d///pcFCxZkvS7k77//ztW+Q2bOnLlHDHXr1gXg1ltv5aeffgJg8uTJ+3SYSk5OpkqVKlSpUoWKFSvyzz//sGLFijzFAOzxfQwpW7YsCxcupE+fPvz9999s3759j8f//vvvrD+CBxLJn5mIiEi8C1W0rFu3jk6dOu3zeOhkaCgZW7VqFW+//TbHHXccdevWpUOHDhQtWjTfcRQtWpS77rqL0qVLc/755/POO+8AZB1PhE6eA9x2222ULVuWVq1a0aBBg6yE8sYbbe7a+PGQraozLPs79njxxReZPXs2GzduzBqBg/CPtapWrcprr71Gamoqa9asIT09PdfbkMSipC2C2rRpw9SpU9m5cyfjxo3LGjUqWbJkVmnkrFmzuO+++/b45csutyWK2UvzDj/88KzbVapU2ee5KSkp+yRy+dl3SOjMEthoX07xrF27dp/XHXnkkVm3S5Qoka8YYM8/YqF/Eh9++GHWKGBO9k7ichLpn5mIxKd/f/PvrNu9LugVYCQiwcvp/3p2GzZsAODKK6/k559/5scff9xjhKhKlSr07Nkz3807KlWqlFXJdMghh2Tdv3PnTmDP46Tsx0ahY5Rly2DaNHjtNWjcOHf7zsjIYMmSJQCUKlWK8uXLs3TpUu688859ToyHhHPckZGRwd13373fUkoddxROBx0Ads6955xb6Zyble2+J5xzS5xzMzIvF2V77EHn3Dzn3Fzn3PnRCjwWtWrVKuus0eDBg/n5558BaNasGSVLlgSsvWvo4L9Xr15MnjyZcePG5XmflStXzrq9cuXKrNurVq3a57nZlx7o378/U6dO5ZNPPslxu865sGPIvi5d9pGy7PHsvXYdsMcZttzsb38+/vjjrNtnnnkmsPs9V6pUicGDB/PTTz/Rq1fuDrYi/TMTkfj0498/Zl1ECrvQ//Xjjjsuq7Qv+2XIkCGAnbh+6aWX+Pbbb3njjTe47777qFy5MqtWrdpjekRejwMONlqX/TgpewK3YsUKNm60EbZrroE77sj9vocPH56VnDVp0oQiRYowefLkrPvuv/9+Jk6cSGpqKocddtg+r9/fe160aFFWwnbRRRcxduxYUlNTc2wyJ4VHOFW7/YALcri/l/f+1MzLVwDOuROAjsCJma95wzmX/7HvOHHYYYfROPM0zS+//JI1HB7qGgm7z/wAHHrooezYsYPXX389z/usW7du1hmmYcOGsWDBAlasWMFHH320z3N37doF2B+J0qVLs3HjRt56660ct1u2bFnA5oZt2rTpgDGccsopWTEMHTqUefPmsXbt2qwSBWCPLk6RlJGRwcKFC3nssceYOXMmACeccELWyGbo+120aFFKly7NqlWr6N+/f47bCnX8XLRo0R5nwiL9MxMREYl3oZOjaWlpDBgwgE2bNrFlyxamTZvGY489lrV22ZgxYxgyZAgbNmygQYMGtGnTJmuUK3ulTm6OO3KjQYMGWbffeeedrOkr06b9wvz5th5b376Qm5xx/fr1DB06NOskcLFixejcuTOw5zFD6dKlSU9PZ8CAAVmNUbILvWeAefPmZd3Ovo2SJUtSvHhxpkyZkjVPTwqng5ZHeu/HOedqhbm9y4CPvffbgb+cc/OAxsCkvIcYX9q0acOkSbvfbunSpTnrrLOyvm7WrBmDBg0CyJpgG5rEmhdly5bl2muv5Z133mH58uV06NABIMczOs2aNSMlJYXt27dz5ZVXHnDfJ554InPnzmXp0qVZC4K/8cYbWUlpdoceeijdunXjf//7HytXrtyjiyNA69ats1rZRtKTTz7Jk08+ucd9derU4X//+19WU5hmzZoxZ84cVq5cyUUX2YDw/t7zCSecwOTJk5k5c2bWz2zo0KER/5mJiIjEuy5dumS1+u/du/c+UxFCzcj+/PPP/S4e3bRp06zbuTnuyI22bdvy/vvvs3jxYoYMGcKQIUPwHhYurERGxhqOOQbKlAlvW9mXIwgpXrw4jz32GPXr1wfsJHWxYsXYtWsXjz/+OI8//jjlypWjbNmybNy4cY/XnnjiiVm3Q8dON998M8nJydSoUYMlS5Zkxeyco1q1aizNvnCcFCr56R7ZzTn3S2b5ZIXM+2oAi7I9Z3HmfftwziU751Kdc6k5lfLFqxYtWlCs2O5cuFmzZnvUWJ9++uk8+OCDVK9enUMOOYTTTz99j/KAvEhOTubGG2/ksMMOy5qIm1Ob/UsuuSSr+UepUqVo1apVjksDgP0xbt68+R5ngQ7kqquu4sUXX6Rhw4aULl2aEiVKUKdOHe6666797iMSihcvTqVKlWjYsCH33HMP/fv332Ne3c0338xVV11F+fLlKVOmDJdddhn33ntvjtu67777aNSo0T4LoEfjZyYi8WfzwnpsWaR1GUXA5rAPHDiQdu3aUbVqVYoVK0bFihU55ZRTuOOOO7JG4ho3bkybNm2oXr06JUuWpHTp0tSpU4fbbruNO7LVJOb2uCNcxYoV4/XXX6dx48aUKFGC6tWrU7PmQ6xbdyJHHw2HH14uV9srUqQIZcqU4ZhjjqFDhw58/PHHXHjhhVmPH3300fzf//0ftWrVokSJEpx44om89tprlMkhM2zTpg3XXHPNHiWcoZhfeuklGjZsyCGHHELNmjV58sknOe200/L2TZCE4HLqerPPk2ykbYT3/qTMr6sCqwEPPAVU897f7Jx7DZjsvR+Y+bx3ga+99wfsTZqUlOSzt1wXERGJNaWPmkP6lnKc9PTlTLttatDhiEiYfv75Z+rWrUuZMmUYMgSuuGI6Vat2o2bNHbRs2ZIXXngh6BBFAHDOTfPe51ielqfukd77rG4Tzrm3gRGZXy4Bjsz21JqZ94mIiMStjAzYvuJoMnaUYv2vzYIOR0Ry4e233yY1NZVDDqnIpEk7KV16PTVq2BSTO/LSgUQkAHkqj3TOVcv2ZTsg1FnyC6Cjc+4Q51xtoC6g05EiIhLXli6FjB2lAFg19sqAoxGR3LCyy2OZPHkLzm3mjDOO4LLLLmXQoEHUrl076PBEwnLQkTbn3EdAC6Cyc24x8DjQwjl3KlYeuQDoAuC9/8059ynwO7AL6Oq9z3lxKxERkTgRWi6pzHGpbPi9KXPnwvHHBxuTiByc97B6dUd++KEjxx8Pw4fDsccGHZVI7h10pM17f433vpr3vrj3vqb3/l3v/Q3e+5O996d47y/13i/L9vxnvPfHeO+P995/Hd3wRUQkVsyfDy+8YAdJiSaUtNXs8DKu6E7eeCPYeETk4HbsgORkuPNOuOgimDRJCZvEr/x0jxQREcnyyivw3//CTz8FHUnkpaVBkRJbKX3kXCo0+o5+/SCCS0mJSIStWAGtWsE778DDD8OwYVAud40iRWKKkjYREYmIlBS7HnzAfsHxKS0Njqq9nYfPeZB77y7Fhg0wcGDQUYlITqZNg6QkmD4dPvkEnn4aiuiIV+KcPsIiIpJvK1bArFngnCVtiVYimZYGp59Snvb12/NAxxY0bAivv55471Mk3n30ETRrZknaxIlw1VVBRyQSGUraREQk38aOtevkZPjrLzvDnSh27rT5esdlrqvtHHTrZknquHHBxiYiJj0dHnwQrr0WTj/dyrRPPTXoqEQiR0mbiIjkW0qKzRfp0QOKFk2sEsm//rIDwlDSBnDNNVCxIrz2WnBxiYhZvx4uuwyeew66dIHvvoPDDw86KpHIUtImIiL5lpICLVrYgVLr1olVIhnqHJk9aStVCm6+GYYOhSVLgolLROCPP+CMM2DUKHjjDejTB0qUCDoqkchT0iYiIvny998wb551agPo0MG+njkz2LgiZe5cu+45twvXD7me64dcD8Dtt0NGBrz1VoDBiRRio0ZB48awerWNrt1+e9ARiUSPkjYREcmX77+361DSdvnliVUimZYGlSrBX9unMWf1HOasngNAnTrQti307WvrQYlIwfAeeva0tdeOOsrmr51zTtBRiUSXkjYREcmXlBSoUgVOPNG+rlLFSiU/+ywxSiTT0vYsjcyua1frnPn55wUbU8hXX8G77wazb5EgbNsGN90E994L7dpZh8hatYKOSiT6lLSJiEieeW9JW8uWe66D1KGDJTuzZgUXW6QcKGk77zw49thgGpIsX26d8rp3tw6XIolu6VIbUfvgA2t69OmncOihQUclUjCUtImISJ7NmweLF+8ujQxp186SuHgvkdy0yQ4U95e0FSlio20TJ8LPPxdsbPfea13zNm+2xYRFEtmUKbZg9u+/WwOgRx/VgtlSuOjjLiIieZaSYtd7J21Vq0Lz5vGftP3xh10ff/z+n9OpE5QubYttF5SUFBg0yNqbw+518kQS0YAB9vekZEmYNMnmzYoUNkraREQkz1JSoGZNKxHcW4cOdlb8998LPq5Iyand/97Kl4frr7ck6p9/oh/T9u1wxx3WCKVXL5tLGGoGI5JovvvO5rA1a2YNR046KeiIRIKhpE1ERPIkI8OStlatwLl9H2/f3u6P59G2UNKWU1KaXdeu1iDhvfeiH1PPnrYMwWuv2XpxLVvC+PGa1yaJZ/16Ww/x+ONhxAjr4ipSWClpExGRPJk1y9ZH2rs0MqRaNTs7Hu9J21FHWXJ0IKecAmefDW++Cenp0Yvnr7/gqafgiivgwgvtvhYtYMsWSE2N3n5FgvDvf9vi9QMGHPx3UCTRKWkTEZE8Cc1na9ly/8/p0AF+/XX3AtXx5kCdI/fWrRvMnw/ffBOdWLyHO++0NfB69dp9f/Pmdq15bZJIvvwS3n8fHnzQFtAWKeyUtImISJ6kpFjZ4FFH7f857dvbdTyOtnmfu6StXTsbXYxW+//hw2HkSHjySTjyyN33V6li83w0r00SxerVcOutNoL92GNBRyMSG4oFHYCIiMSfXbvghx+gY8cDP69mTTjzTEvaHn64YGKLlNWrYd263UnbS+e/dMDnFy9u3RyfeMKWQjjYPLjc2LQJ7roLTj7ZrvfWsqUtsr1jB5QoEbn9igSha1dr6jNqlD7PIiEaaRMRkVybPh02bNj/fLbsOnSAGTMskYkne3eObH5086zL/iQnQ7Fi8MYbkY2lRw9YtMjmzBUvvu/jmtcmieKTT2zR7CeegAYNgo5GJHYoaRMRkVwLzWdr0eLgz73iCruOtxLJ0Dy8cMsjwcojr7jCukhu3hyZOGbNsjlsN98MZ52V83M0r00SwbJltpxF48bw3/8GHY1IbFHSJiIiuZaSYvOoqlY9+HOPOgqaNIm/pC0tzUa1jj46d6/r1s1alX/4Yf5j8N4OYsuVg+ef3//zKle20kklbRKvvLeR6i1boH9/G7EWkd2UtImISK5s327rgrVuHf5rOnSAadOsu2K8SEuDY47J/cHjWWdZWddrr9mBaH707w8//mgJW+XKB35uy5YwYYLNaxOJN/362Vpszz0H9eoFHY1I7FHSJiIiuTJlCmzdGt58tpBQieTnn0cnpmjYu3PkBQMvyLociHPWSOGXXyy5zat//oH77oOmTa008mBC89p++inv+xQJwsKF0L07nHOOLWshIvtS0iYiIrmSkgJFiuyeRxWO2rUhKSl+SiTT061xSvakbfWW1VmXg7n2WihfHl5/Pe8xPPggrF0LffrY9/tgNK9N4lFGhp2U8N7WZQvnsy5SGOlXQ0REciUlBRo1sqQkNzp0gKlT7ax6rFu0yMpAjz8+b68/9FA7EP38c1i6NPevnzwZ+va10YdTTgnvNZUq2XOVtEk8efNN+5vy0kt2ckdEcqakTUREwrZ5syUUuSmNDOnQwa7joURy73b/eXH77baeXd++uXvdrl1w221Qo4a1Pc+N0Ly27dtz9zqRIPzxh3WJvOACuOWWoKMRiW1K2kREJGwTJsDOnXlL2o45Bho2jI8SyUgkbcceCxdeCG+9lbvmIK+/DjNnwssvQ9myudtnixY231Dz2iTWpadDp062ePY779hcUBHZPyVtIiIStpQUa4O/v/XCDqZDB5g0ycoPY1lamiVM4SxpcCDdusHy5TB0aHjPX7oUHn3URh5CzVtyo3lzO/hViaTEup49YeJE67Jao0bQ0YjEPiVtIiIStjFj4IwzbM5WXoRKJIcMiVxM0RDqHJnfs/8XXAB16tiBaTj+/W8blXv11bztu2JFzWuT2Ddrlp2caN/emvaIyMEpaRMRkbCsXQvTp+etNDLkuOMsqYj1Esm92/3nVZEitjj2+PFW8nggo0fDp5/CQw9ZaWVeaV6bxLKdO+HGG+Gww6wzqsoiRcKjpE1ERMIybpy1585P0gY22jZhQt66KhaE7dthwYLIJG0A//oXlCp14Pb/27bZ2m5168L99+dvfy1a2PamTs3fdkSi4Zln4OefrUFPlSpBRyMSP5S0iYhIWFJSLPlo0iR/2+nQwdZkitUSyT//tPgilbRVrGglYIMG2WhlTp5/3taFe/11OOSQ/O3v7LM1r01i07Rp8PTTcMMNcPnlQUcjEl+UtImISFhSUqBZs/wnFfXrw4knxm6JZCQ6R+6ta1fYsgX69dv3sXnz4P/+D66+Gtq0yf++KlaEBg2UtEls2bbNyiKPOAJeeSXoaETiT7GgAxARkdi3YoU1D7juushsr0MH6NHDOisecURkthkpc+fadd26e97/QbsP8rzNhg2t4+brr9uC2UUyT5l6bx0mS5SwxYUjpWVLW7R4+/b8J9kikfDYY/D77/DNN1ChQtDRiMQfjbSJiMhBhUZt8jufLSRUIhluK/yClJZmrf4PO2zP++tXqZ91yYuuXa30cvTo3fcNHgyjRlnJWPXq+Qh6L6F5bVOmRG6bInk1YQK8+CJ06QLnnx90NCLxSUmbiIgcVEoKlCsHp50Wme2deCLUqxebJZKR6hy5tyuusGQw1P5/40a4+2449VTrMBlJmtcmsWLzZrjpJqhVC154IehoROKXkjYRETmolBQ45xwoFqGieudstG3sWFi5MjLbjJRoJW0lSthIw1dfwfz58PjjsGyZtT2P1Pc1pEIFSwaVtEnQ7r/fPu/vv28L1otI3hw0aXPOveecW+mcm5XtvorOuW+dc39kXlfIvN8553o75+Y5535xzkXonKyIiATl77+tWUbr1pHdbocOtoTAsGGR3W5+rFtnSWQ0kjaA5GSbz3bXXdC7t32d326c+9OyJUycaGWSIkH47jubx3n33XbSR0TyLpyRtn7ABXvd9wAwxntfFxiT+TXAhUDdzEsy8GZkwhQRkaB8/71dR2o+W8gpp1izj1gqkfzjD7s+/vh9H0vqm5R1yasaNaB9exg50ro8Pvtsnjd1UC1aWCMSzWuTIKxfb2sU1qtna7OJSP4cNGnz3o8D/tnr7suA/pm3+wOXZ7t/gDeTgfLOuWoRilVERAKQkmKL4J54YmS3GyqRTEmB1asju+28ika7/711727vvWdPS9yiRfPaJEj33mvlv/372/qOIpI/eZ3TVtV7vyzz9nKgaubtGsCibM9bnHnfPpxzyc65VOdc6qpVq/IYhoiIRJP3llS1bLm7TX0kdegA6ekwfHjkt50XaWn2PuvUid4+zjrLllC44Ybo7QOgfHlbakBJmxS08ePhnXfgnnugceOgoxFJDPn+F+y994DPw+v6eu+TvPdJVapUyW8YIiISBfPmweLFkS+NDGnY0BKkWCmRTEuzLnfRXtusoP7ttWwJkyZpXpsUnJ074bbb4KijrNmOiERGXpO2FaGyx8zrUO+vJcCR2Z5XM/M+ERGJQykpdh2tpC1UIvndd/DP3oX4AYhW58ighOa1TZ4cdCRSWPTqBb/9ZktbHHpo0NGIJI68Jm1fADdl3r4JGJ7t/hszu0ieAazPVkYpIiJxZswYqFkTjj02evvo0AF27YIvvojePsLhfeIlbc2aWbmnSiSlICxcCE8+CZddBpdcEnQ0IoklnJb/HwGTgOOdc4udc52B54A2zrk/gHMzvwb4CpgPzAPeBiK8XKjEmnnzrF595sygIxGRSMvIsM6RrVrZiFi0JCXB0UcHXyK5fDls2pRYSZvmtUlBuusuu+7dO9g4RBLRQZfz9N5fs5+H9lmxJ3N+W9f8BiXx44MP4KefrIV1aqot6CoiiWHWLOvqGK3SyJBQiWTv3rZOWvny0d3f/hRE58ggtGxp39utW9XFT6Jn+HAbLX/hBZvPJiKRFYVeYFKYDB1qTQQWLYIbb7Qz8yKSGELz2Vq2jP6+OnSwBgZffhn9fe3P3Ll2nWhJW4sWsGOH5rVJ9GzaBHfeCSefbEtaiEjkKWmTPPvzT/j1V+jWzSYejxgR3YViRaRgpaTYXLaCOGvepAkceWSwJZJpadY18sgjD/7ceKJ5bRJtTz5pJ2/79IHixYOORiQxKWmTPAutq3T55XDHHXD99fDYYzBqVKBhiUgE7NoFP/wQ/dLIEOfgiivs78eGDQWzz72lpUHdutFZjy5Ihx0Gp52mpE2i45df7MTtLbfAmWcGHY1I4kqwf01SkIYNgwYNoHZtO+B66y046SS49lpYsCDo6EQkP6ZPt+SpoJI2sBLJ7dtt1D4IB+sc+fV1X2dd4k3LllYeuXVr0JFIIsnIgNtvt/nszz138OeLSN4paZM8WbkSJkywUbaQ0qVhyBBIT7eDLy3mKhK/QvPZWrQouH02bQrVqwdTIrlrl5V8Hyhpq3JolaxLvAnNa5s0KehIJJG89x5MnAgvvgiVKgUdjUhiU9ImeTJihJ1hy560gc1/GTAApk3b3fpXROJPSoqNnFetWnD7LFLESiS//toaGxSkBQsscTv++ILdb0HRvDaJtFWr4L//hXPOsUZkIhJdStokT4YNs3WVGjTY97FLL4WHHoK334Z33y3w0EQkn7Zvh/HjC7Y0MiQ0Sj9yZMHuN1Hb/YeUKweNGilpk8i57z47ufLmm9Fdx1FEjJI2ybVNm2D0aBtl298f6h494NxzoWtXG3UTkfgxZYrNfWq9z2qc0XfWWXDEEQVfIhlO0rZq86qsSzwKzWvbsiXoSCTejR0L/fvDvfdC/fpBRyNSOChpk1wbPdrOxO9dGpld0aLw4Ydw+OF25nzNmgILT0TyKSXFSumaNy/4fRctCu3b20jb5s0Ft9+0NGumcKB5ORcOujDrEo9atLC18DSvTfJjxw7rGF2rFjzySNDRiBQeStok14YNg4oVbY7EgVSpYmfLly615QDS0wskPBHJp5QUK6UrXz6Y/XfoYCN9Xxdgk8ZQ58hELvM66yxLilUiKfnRsyfMng2vv24NyESkYChpk1zZuRO+/BIuuQSKFTv48xs3ht694Ztv4Kmnoh+fiOTP5s1WQhfEfLaQ5s3tpM+QIQW3z4O1+08Emtcm+TV/vk1/uOIKuOiioKMRKVyUtEmujBsH69YduDRyb8nJ0KkTPPkkfPVVlAITkYiYMMFOzgSZtBUtageE33xjHR2jbcsWWLQo8ZM2sHltU6ZoXpvknvfQrZudsH355aCjESl8lLRJrgwbBqVKwXnnhf8a5+CNN+DUU+G66+xMnYjEpjFjoHhxK6ULUtu2sHatjfpF27x5dl0YkrbQvLaJE4OOROLNkCFWsvzUU1CzZtDRiBQ+StokbN5b0nbeebmvYy9VCj7/3G5fcYXNVxGR2JOSAmecAYceGmwcbdrYiFtBjM4nerv/7DSvTfJi40Zbe/XUU220TUQKnpI2Cdv06bB4MbRrl7fX16kDAwfCjBnWecr7iIYnIvm0dq39ngdZGhlSvrw1OyqI9drmzrXrY4+N/r6CVrYsJCUpaZPceewxWLYM+vQJbz67iESekjYJ27Bh1gb84ovzvo22be2Pf79+tvi2SGEzaZK1zI5F48ZBRkZsJG1gfy9++cXmm0VTWhrUqAFlykR3P7GiZUuYOrVgl1SQ+PXzz9ZQrEsXaNIk6GhECi8lbRK2YcOsq9uB1jEKx2OPwfnnw513wk8/RSQ0kbiQmgpnngkPPBB0JDlLSbFS5lg5MGvb1q6jXSJZGDpHZqd5bRKu9HS47TaoXBmefTboaEQKNyVtEpZ582DWrNx1jdyfokVh0CCoVs3mt61enf9tisSDfv3sundvmDkz0FBylJJiJYmHHBJ0JKZ+fTj6aCVtkaZ5bRKut9+2UdmXXrLF50UkOEraJCzDh9v1ZZdFZnuVKlljkpUr4ZprtPC2JL7t2+HDD62VfcWKcPvtVooYK1assBMzsVIaCdZ5tm1b+O472LYtOvtYswb++adwJW1lysDppytpkwNbscKqAlq1gmuvDToaEVHSJmEZOtS6RtWqFbltNmoEr79uB2SPPRa57Up4du60LmBffhl0JIXDiBHW6KN7d3jhBZvb9t57QUe1W+gAPpaSNrCkbcsW+OGH6Gw/1Dny+OMP/tzU5NSsS7wLzWvbtCnoSCS/vLf1UE8+GTp3hr59bSQ/v2sc/uc/1un5jTfsBIqIBEtJmxzUihU29yESpZF769wZbrnFauW/+CLy25f9u/deS5rvuksjnQWhXz9rdtG6Ndx4o80Pvf/+2CkPTkmBcuXgtNOCjmRPLVpAyZLRK5EsTO3+s2vRwg7qNa8t/n36qZUxlihhc8+7dLGTrOXL28/5/vttjbWlS8Pf5pgxNo3h/vvDO6EhItGnpE0O6ssv7UxeNJI2gFdftVG3G27YvcitRFe/fjavqkkTWLBAo23RtmKFLUp7ww02lyi04PyGDfDf/wYdnUlJgXPOib123qVL2+jfyJHRWSYkLc3ecySrCOLBmWfa+1aJZHxbsQK6doXGjWHKFDsJ9McftrzOzTfbKHWvXjZ/vEYNOPJI6NABXnwRfvzRHt/b9u22LM8xx8CDDxb8exKRnMXYv2eJRcOG2QHNKadEZ/slS8LgwZa4XXEFTJ5sHewkOqZOtW5grVvbgfBxx8Err0QvKRc7Y52eDjfdtPu+E0+08qPnn7eDq2bNgovv77/thEnXrsHFcCBt29pIW1pa5M/6p6XZGpLFi0d2u7FO89rin/c2N3bTJjsRFzrhcuyxdrnuOvt62zZbH3XKFPv/OmWKzSkHO4l08slwxhl2Eq9JExu5S0uDb77R/2KRWKKRNjmgjRttztnll0e3pr1WLTuw/eUXm2cl0bFsmS2OXr06fPKJdQns1s0O3H75JejoEpP3dkDVpAnUq7fnY48+CkcdZQdeO3cGEh4A339v161bBxfDgVx0kV1HY6Ht3HSOnL1qdtYlEbRoYcuuaF5bfPr4Y5tv/tRT1ml1f0qWtKSse3f46COYP99G6L780hqNVK5sTZL+9S844QR44gm4+mpbmkdEYoeSNjmgUaOsVKJdu+jv64IL4OGHrTlDqDW6RM727TaSuW6djZ6G1tu75RYrQXvllSCjS1wzZsCvv0KnTvs+duihVqY6a1aw3/+UFKhSxUb/YlGtWnYwGel5bRkZVkoWbtJ2w9Absi6JoGVLm9c2YULQkUhuLV9uJ9zOOAPuuSf3rz/8cLj4Ynj6afj2W2uS9Ntv9v/3v/+1v0siEluUtMkBDRtmZ+HOPLNg9vfkk3YgcccddqArkeG9/YOfNMkS4uylrhUqWGOMQYNg1arAQkxY/frZiObVV+f8+GWXwSWX2NntRYsKMjKzfbuNprdsCUVi+D9C27YwbpyN/kfKkiXWHa+wNSEJ0by23FuxwpKkDz8MLgbvrcR982Z4/30rccyvIkXsxMi//mUl24cfnv9tikhkxfC/aAnazp3WpvySSwquOUHRovbP8LDDbLL0hg0Fs99E16cPvPMOPPQQXHnlvo/fdZcdvL/1VsHHlsh27LBk+LLLDrwwbe/eNurTvXvBxQa2z5tusq5y119fsPvOrbZt7W/St99GbpuFtXNkyKGHWgMLJW0Ht22bJTN161pjjxtuiP6i7/vz4Ye2duozz+xbci0iiUtJm+zXDz/A+vUF36DiiCOsVn/ePLj11uh0jCtMxo2zpKxtW5v7kJP69W3+whtvWKIhkfHVV7Z4c06lkdnVqmVrFQ4dGp15Wznx3hqhfPKJHYxecknB7DevzjzTTuZE8kB57ly7LqxJG+ye1xbJEcxE4j189pmNQj3wgH2/pk2zlvpXX23lzwVp2TK48077fbj77oLdt4gES0mb7NewYTbXqU2bgt/3OefYWcRPP7W1xCRv/v7bRiyPOcZGfA5U/ta9ux0QDB5ccPElun797CREOL9D99xjyXO3bjm34Y60nj3h5Zft537ffdHfX34VLw7nnWdJW6RO5KSl2d+46tUjs7141LKldTbVvLZ9pabaeopXXWXdNr/7ztYTPe00a+JRoYLNC1u8uGDi8d7WYNu6NXJlkSISP5S0SY4yMixpO//84Fr+/ve/9g/xnnusTb3kzpYt1kBm+3YrpTnssAM///zzd7f/l/xbtcpGzW64Ibzy4hIlbKRzwQJbbD6aBg60RO2qq+Cll6LbGTaS2ra1Ews//xyZ7YU6R8bL+4+Gpk0tIVaJ5G5LlljZ8Omn22ekb1/7zGXvrlq9uv1+b9hg/6cKYqRy4EBLFp99tnCPDosUVkraJEfTptk/riDX7ipSBPr3t3+OV10F//wTXCzxxnsrLf35ZxthC2dtqyJFrIxy6lRby0fy58MPrTNf9rXZDqZFC0vy/vc/mDMnOnGNHm3NBlq2hAEDYrv5yN4uuMCuI1VCmpt2/4lK89p227zZmmEdd5yV6N9/v3UXvfXWnEe1Tj7ZSidnzbJSyV27ohfb0qX29/mss+xaRAqfOPp3LQVp2DD7J9W2bbBxVKxoJZJLl1qHw4yMYOOJFy+9ZEnDU0/ZWeBw3XSTjchptC3/+vWDpKTct9F/4QU7kL7jjsjP50xNhfbtLaahQ62rZTypWtVGPyIxr23HDvjrLyVtYCcLUlML77y2jAz44AM7ufXEE/Z/b84ceO45KFfuwK8NzQX++mtLpqIxB9t7SE62qgmVRYoUXkraJEfDhlktf2gtryA1bmxJyMiRNgIhBzZ6tJWWduhg3SJzo0wZ6NzZ5rUtWRKd+AqDmTOtQcHBGpDkpGpVK3/6/ntbCDdS5s2zRaorV7YDzIOVy8aqtm1hypT8L08xf74drIczCp3oQvPaxo8POpKCN2GCrXV2441QrRr8+KOdKKxdO/xtJCfb39w337TOkpE2YID9/3v2WeteKSKFk5I22UdaGvz+e7ClkXvr2tVa1T/8sHW1lJz9+Sd07GgjKe+/n7e5Ot262cHsG29EPr7Con9/myfUsWPeXp+cbCNK99xji6Hn14oVVlqYkQGjRtnBabxq29ZGHr75Jn/bKezt/rMrjPPaFiywksZmzaySY8AAOxnQrFnetvd//2f/o+6910axI2XJEmsWdPbZKosUKeyUtMk+hg+361hK2pyzdcaOOQauucYOQmVPGzfaemDO2UhpmTJ5207t2nDppbZm29atEQ2xUNi50+YRXnpp3keqixa1tfVWrYJHHslfPBs37m7gMXJk/I8snXaajUbmt0QylLTlZuSicunKWZdEUro0NGlSOJK2DRvgwQdtfbMvv4THH7elH264IX/zO0NzsJs0geuui0zzrNDc5B074L334mv+qYhEnv4EyD6GDbMDo6OOCjqSPZUrZ2V7a9fCtddaOY+Y0CLJs2fbult16uRve9272/pigwZFJr7C5JtvYOXK3DUgyclpp9kI8xtv2HyjvNixw8pkZ8ywkq8mTfIXUywoUgQuvNC+z/lp/JCWBlWqHHjR8719c/03WZdEE1p/bMOGoCOJjvR0ePttS9Kfe86aW6Wl2Ry2Qw+NzD5KlbKTnkccYese/vVX/rbXr5+VMj/3HBx7bERCFJE4pqRN9rB8OUyaFFujbNmdcoodxKak2D9bMc88YyU5L74I556b/+2dc459r195JX4XN9+1y7pgFnT8/frB4Yfv7nSYH089ZaNKt92W+5MUGRlw8802x/Htt4NvKhRJbdta2eikSXnfhjpH7ik0r23AgPj9nd+f0aOhYUMrO65b10bBBgyAmjUjv6/DD7dR4J07d39O82LxYls8u3lzK1kXEclX0uacW+Cc+9U5N8M5l5p5X0Xn3LfOuT8yr3NxHlOC9uWX9g87VpM2sHbl//oXPP10/ue1JIIvvoDHHrPynrvvjsw2nbNtzZplDTHizdixdpDWtGl0GgPsz5o19jt0/fU2Ryi/DjvMmvBMm2blqrnxwAM2UvrMM/b7kkjatLG17/LT+l9J257OPBMaNYI777SOiNFacqIgzZplo7Lnnw+bNtlo848/2nzRaKpXD4YMseY/V1xhI9654T3ccoudeFJZpIiEROJPQUvv/ane+6TMrx8Axnjv6wJjMr+WODF0qJXWnXRS0JEc2Guv2Ro5118PixYFHU1wZs+270FSkh3UR3KR4GuusfKxeGr/v3ixNf9o2dLmcjVpYnPC5s8vmP1/9JGdYc9vaWR2HTvaor4PPWQj4eF4+WVbOqBrV5u/k2gOO8waRuR1XtvGjTbHT0nbbiVL2sj0K6/YSNTJJ1tTjXgsl1y+3EbVGjSw99Szp/2tvPLKgltIvUULm4edkgJduuRu9PK996xh0PPP2zxuEREAvPd5vgALgMp73TcXqJZ5uxow92DbadSokZfgrV/vfYkS3t9zT9CRhGfuXO/LlvX+jDO837496GgK3tq13tet6/3hh3u/aFF09vHII9475/28edHZfqRs2+b9s896X7q09yVLev/EE95v2eL933/bZ6RVK+8zMqIfR1KS9w0bRn67c+fa7+Z11x38uR995D14f8UV3u/aFflYYsULL9j7XLgw96+dNs1e+/nnuXvdDwt+yLokshUrvO/c2X73q1b1vl8/79PTg47q4DZv9r5HD+8PPdT7YsW8797d+9Wrg43p8cfts/bUU+E9f+FC78uV875Fi/j4notIZAGpfn951/4eCOcC/AVMB6YByZn3rcv2uMv+9V6vTQZSgdSjjjqqQL4RcmCffGKfiHHjgo4kfJ9+ajHffXfQkRSsXbu8v+AC74sX9/7HH6O3nyVL7OAnlr+/I0Z4f+yx9jlo1877v/7a8/E33rDH3n03unH8+qvt5+WXo7P9Rx+17Y8Zs//njBljn4nmzb3fujU6ccSK33+378ebb+b+taHE9tdfc/e6Rm81yroUBlOnet+4sX2vmjb1PjU16Ihylp5uiWWNGhZr+/bep6UFHZXJyPD+hhssrkGDDv7cNm0s6fzzz4KJT0Riy4GStvyWRzbz3p8GXAh0dc4132sUzwM5FgV47/t675O890lVqlTJZxgSCcOG2cK7Z54ZdCThu/JKm4Px8svw+edBR1NwHnnE5vO9+mre1xUKR/Xq1mXtvfespCyWzJtnHdouvtjmN40ebfNIatXa83ldutgaR/fcYyVx0dK/v8Vx7bXR2f6DD1rp8h135DxHZsYMm4t6/PHWwa5kyejEESvq1bOfdV5KJOfOtTI5lZ4d2OmnW7OX99+3NSBPP93KDvO7sHkkpaTYXLxOnezv1bhx9r8gVhahds4aAZ1zjs0t/fHH/T/3nXfg22/hf//LfwdgEUk8+UravPdLMq9XAkOBxsAK51w1gMzrlfkNUqJvxw6b1H/ppbZGVDx58UVo3Ng65c2bF3Q00eM9TJxoSchzz9l1ly7R32/37javpV+/6O8rHJs32yLrJ55oDUdeeAFmzrTmFDkpUsQOhrZti14Xtl27YOBASyCjdQ6qVCmbyzl3rn3ms/vrL2u4UL68JfPly0cnhljinHXnGzPGfra5kZZmS5qUKhWd2BJJkSKWEKWlWXOi996zuYCvvpq/JRfya84cO2nTurUtA/PhhzZ/7eyzg4tpfw45xE4o1a5tJ1ZCawRmt3Ah/Oc/Nh/3ttsKPEQRiQN5Ttqcc4c658qGbgPnAbOAL4DQNPybgOH5DVKib+xYOzCP5a6R+1OihHUFK1rU1qRKtAWh58yBRx+1UYGzzoIPPrAEtXfvgtl/48Zwxhm2v4yMgtlnTry3n3O9evDss3D11Xbwc++99hk4kOOOsyUihgyJzojs6NHW/CCSDUhycuGF1o3uqad2rwG1apV1x9u+3ZoX1KgR3RhiSdu2sGVL7heFVufI3At1Mv3lFxvZuusuW0uwoBfkXrXKGuycdJKNqj33nP2NvOaa2O6yWLGinRgtUsQ+t6tX734s1C0yIwPefTe234eIBCc/fxqqAuOdczOBqcBI7/03wHNAG+fcH8C5mV9LjBs2zBYYjcQaX0E4+mhLZmbOtIOJeLdsmbWqT0qC+vUtSalb10rwVqywf+wHS1QiqXt3G8XMa7e+/Jo1y86oX321lfCOH2/rLFWrFv42/vMfWwaga1f455/Ixte/v8V10UWR3W5OXn7ZyjDvvNNGHS++2Dqojhhhn5XCpEULGy3LTet/75W05ccJJ1gJ3+ef24m+li3t9zLaXXy3brUE7ZhjrFPubbfZ36T774+fUuBjjrElWhYtshOkoRHivn3hu+9sBL127UBDFJFYtr/JbgV5UffIYKWne1+9unWbi3cPPGATvvv3DzqS3NuwweJu08b7IkXsfTRq5H2vXt4vXRpsbDt22CT/c88t2P2uXWsd4IoW9b5iRWs6kZ+OiNOn27b+9a9IRej9P/9YZ8e77orcNg+mZ0/7fJxwgn1Whg0ruH3HmrZtva9TJ/zuoMuX2/fulVdyv6/C1ojkYDZvtu6IJUta59ann458A5z0dO8HDvT+qKPs53bJJd7Pnh3ZfRS0UAOtjh29nz/f+zJlvG/dumA63IpIbOMAjUiKBZ00Svh27rT1c5o2jWz5RGoqLF0an6WRe3vqKZs4f9ttVjZ13XVQtmzQUe3fzp1W0jZwoJ2B3brVzrQ+9JDFXq9e0BGa4sWtAcbDD8Nvv9l8smjKyLDRqwcesHKoLl1sMfVKlfK33YYN4b777Iz9Ndfsfx5cbnz8sc0J7dQp/9sK15132hzDX3+1UYfLLiu4fceatm1tpG3u3PB+X0LziY4/PrpxFQalS1vZcadONpL9yCM25+3ll20EOPuaaLt22cjc+vV2yX5770v2x5YutVLg006zz3zLlsG810i68kr7G/TAAzbCBlY9UVBryIlIfHKW1AUrKSnJp6amBh1GzHvoIfi//7MFg195xa4j4cEHrZnDqlVQoUJkthmkZctsgvq0aVbyee211vEsKengry0I3ltiOWgQfPIJrFljycjVV1ui1rRpbP7zXr0ajjwSbrzREoVomT7dEsQpU6yT6WuvWbIVKVu3wqmnWqL1669Qpkz+tnfGGXaCYObMgv25LVhgCXTbtgW3z1i0cKF1kXzxRUscDubdd23+0Pz5uS9FS+q7+49IarL+Z+3t22+tPH3OHDuxk5GxO/navPngry9RwubOhS7lyllTncsvh+uvT6y5Xt7byai334Y+fQqmqZSIxD7n3DTvfY5HrEra4sTy5VYPf8opdrAWanrwf/+Xu3k9Oalf35oXhM74JQLvbVTyrbdsJGTrVjtTm5xsSVxBj755D7Nnw0cfWZez+fNtHsbll1uidv75NpoV62691ZLNxYttYn0keQ+vv26t+StVsrbX118fnUToxx+heXPrhterV963M2eO/f707GlxSzBOOgkOP9zavx/M/ffbSNCWLbnvlKuk7eB27rQTLaNG2d/ZcuX2TMT2vmR/PF7mpkVKero1djn11Ng8USciBU9JWwLo3t0OaOfMgapV4Zln7GCzRAkrWfv3v62tcG6FDjpffTV67dCDtn69lR++9dbukZXQ6FujRtHb77Jl1o78u+/sevFiO1PcurUlI+3axXbpZk5+/dVOHDz3nB38RsqmTZYQfvyxlVUNGBD9Ud877rAz3BMn2mhZXoRGqZcssd9LCcb991tnwzVrLAk4kHbtrETyt99yvx8lbSIiEk0HStoSqNggcS1caAeXN98Mxx5rB/rPPWcHHa1b24HjiSfagrq5zcGHZy7IkMhzYg47zDoGzpxp6/hceaV1mkxKskvfvpFZOHrDBvjyS0uwTzrJFnq94Qa7r2lT+xkuXmzt4W+8Mf4SNoCTT4ZWrewEQqTWaJo925YV+PRT65I5fHjBlOk+95yNMHfubO3ycys93ZLLCy9Uwha0tm3t8/jttwd/bn46R9arXC/rIiIiUpA00hYHOne2krQ//rA5RXv79lsr8/r9d2us8PLL1pY5HE2bWjlLYfv2r1tn39O9R9+6dLEyynDs2GFJ4Hff2WXqVDuQL1XKFng991y7NGiQWHMxvvjCkvxPP7UEOD8+/dRORpQubaWjrVtHJsZwjRxpI3uPP24NFXJj1Ci44AIYPNjWTpPg7Nxpi5q3b2+NMPYnPd0+a3ffDc8/X2DhiYiIhEUjbXFs7lzronfHHTknbGCJ2owZ1pzkp5+sfK17d1i79sDbXrbMko5E6BqZW+XL7x59mzTJFuX+4AMrlzz9dJscvmnTnq/JyLDn9+xpoysVKsA551ipakaGdQL7/nv7vo8aZV0KGzZMrIQNbFSjTh37vOXVjh124Hz11ZbU/vxzwSdsYO/l2mtthG/WrNy9tn9/+wxcfHF0YpPwFS9u80K//vrAC8D//bd99rRGm4iIxJsEO5xMPI8/bpOzH3jgwM8rXty6dv3xh83Veu01W4z5zTft7HJOvvjCrgtj0hbinM1nev99ay396qu24GlysjV4ue02K2vs2BGOOMImjN97rzWDuflmW5R8zRpLfp9+2hb7zcvcwnhStKi1nJ8wIW8jtIsX2/fplVcscRs71soUg/Lyy1ZC27nz/n9X9rZ+PQwdaglfov+840Xbttag6eef9/+cULt/JW0iIhJvlLTFsBkzrCX8v/9tndHCUbkyvPGGtU0/6SQboTvtNDsw3tuwYTZHLtprbsWL8uWtGcsvv1hziiuusNGU22+3boMXXGBfL15s87BefdXKBMuXDzrygvevf1lJaW5H28aMsc/jr7/aZ7tXr+C7ZlapAr17W3lr797hvebTTy25v+mm6MYm4bvgAjsJM3Lk/p8zd65dK2kTEZF4ozltMeySS2D8eFtYNC+Jgffw+ec2MrRwoZUAvvCCrWm0YYMleN27232Ss3XrbP26Y49VS+a93XWXjUIuXHjwZScyMqzxx6OP2qLGn39uXUtjhff2+5aSYmWSdeoc+PlnnWWfjVmz9LmIJaG1K6dMyfnxbt2sDHrdurz93IbMHpJ1u3399rnfgIiIyAFoTlscmjgRRoywVtZ5HclxzhK12bPhqafgq6/sQPmxx6x5ws6dhbs0Mhzly1uZqQ7M93Xnndaxr0+fAz9v7VobkXz4YZvDNnVqbCVsYD/fN9+EYsWsNPZA57LS0uz3s1MnfS5iTdu2Nq935cqcHw91jszrz+3ZH5/NuoiIiBQkJW0xyHs7wK1a1Q6M86tUKXjkESsNat/eErjOna3kMq/rU4nUrWsHyX367L9l/vTp1txl1CgrJx00yMoqY9GRR9qC3mPG2BzH/RkwwJrLXH99wcUm4Wnb1v5+fvNNzo/np92/iIhIkJS0xaAxY2wO2sMPw6GHRm67NWvaQfP48dC8uSWERYtGbvtS+HTvbqMaH3+872PvvgtnnmkjuuPGWWlarI9MJSfb78Z//mPdVfeWkWFJ2/nnH7wkVApew4Z2siuneW1bt1r3SCVtIiISj5S0xRjv4aGH4Kij7AAyGs46C374wUbfRPKjdWtrZPPKK7tLCrdutc6at9xi69VNnx4/I7pFithyD1u3WpK5t++/h0WL1IAkVhUpAhddZCO7ey/+/uef9hk9/vhgYhMREckPJW0xZvhwm5PxxBNqJS6xzzlrSPLzzzaCO2+eLdj+/vvWdOSbb6w7Yzw57jh48kkYMsQapmTXr58tD3DZZYGEJmFo29aWZJg4cc/71e5fRETimZK2GJKevru73g03BB2NSHiuvx4qVrRSyaQkK0EbMQJ69Ijf8tv//MdK7bp23b1I/YYNlsRdc42tnSixqU0bayizd4lkKGmrW7fgYxIREckvJW0x5OOPrYV4jx520CESD0qXhltvtdG2Y4+1csi2bYOOKn+KFbM5eatXWwIH1nF161aVRsa6cuWsLDenpK1aNShbNpi4RERE8kNJW4zYudNa8Z96qrXpF4knDz9sJZHjx9s6gImgYUP473/tfX37rZVGHn/87rXAJHa1bQu//WZrCIaoc6SIiMQzJW0x4v33Yf58ePppm0wvEk/KlrV1yxKtbPCxx+xA/6ab4Mcf7TrWO2DK7pHer77afZ+SNhERiWdKD2LA1q1WEnnmmdb5TERiQ8mS8M471v7fOc01jRfHHw+1a+8ukVy7FlatUtImIiLxSzOnYsCbb8KSJTBwoM7ii8Sas8+2bq5r1thahxL7nLPRtnfftZNif/xh9+c3aTv7qLPzH5yIiEgeOB9aXClASUlJPjU1NegwArFxI9SpY/NnRo8OOhoRkcTwzTdw4YVWIrlmjY2Szp4N9eoFHZmIiEjOnHPTvPdJOT2mkbaAvfyydah75pmgIxERSRwtWkCpUlYiWaGCzRWuUyfoqERERPJGc9oC9M8/8OKLcPnlcPrpQUcjIpI4SpaE1q0taZs71+a4lSgRdFQiIiJ5o6QtQP/7n5VHPvVU0JGIiCSetm1hwQL47js1IRERkfim8siALFsGvXvDddfBSScFHY2ISOIJdeNduzYySVvfaX2zbic3Ss7/BkVERMKkpC0gzzxjC2o/8UTQkYiIJKajjrKTYrNm2TIA+aWkTUREgqLyyAAsWAB9+0LnznDMMUFHIyKSuEILbas8UkRE4plG2gLw5JPWyezRR4OOREQksXXuDL/9pmZPIiIS35S0FbDZs2HAALj7bqhRI+hoREQSW9268OWXQUchIiKSPyqPLGCPPw6lS8MDDwQdiYiIiIiIxAMlbQVo+nT47DO45x6oUiXoaEREREREJB6oPHI/xo+H1FRo1AgaNoQyZfK/zUcegQoVLGkTEREREREJh5K2/fj6a3j2WbvtnLWLbtRo96VhQyhbNvztjR9v23z+eTjssOjELCIiIiIiiUdJ23488wzceSdMm7b7MnYsDBpkjztnLaT3TuTKldt3W97DQw/BEUdAt24F+jZERERERCTOKWk7gCOOsDV+Quv8AKxYsWciN24cfPjh7sfr1t0zkTvtNJg8GX78EV57zZqQiIiIiIiIhEtJWy5VrQoXXWSXkJUr90zkJkyAjz/e/XipUlCrFtx6a4GHKyIiEdKuXrugQxARkULKee+js2HnLgBeAYoC73jvn9vfc5OSknxqampU4gjKypXWLXLaNPjlF7j5Zjj//KCjEhERERGRWOScm+a9T8rpsaiMtDnnigKvA22AxcBPzrkvvPe/R2N/sejww+GCC+wiIiIiIiKSV9Fap60xMM97P997vwP4GLgsSvsSERERERFJWNFK2moAi7J9vTjzvizOuWTnXKpzLnXVqlVRCkNERERERCS+BdaIxHvfF+gLNqctqDhERETC8cy4Z7JuP9z84QAjERGRwiZaSdsS4MhsX9fMvE9ERCQuDZ0zNOu2kjYRESlI0SqP/Amo65yr7ZwrAXQEvojSvkRERERERBJWVEbavPe7nHPdgFFYy//3vPe/RWNfIiIiIiIiiSxqc9q8918BX0Vr+yIiIiIiIoVBtMojRUREREREJAKUtImIiIiIiMQwJW0iIiIiIiIxTEmbiIiIiIhIDFPSJiIiIiIiEsOUtImIiIiIiMQw570POgacc6uAhUHHkYPKwOqggxCJMn3OJdHpMy6FgT7nUhgk+uf8aO99lZweiImkLVY551K990lBxyESTfqcS6LTZ1wKA33OpTAozJ9zlUeKiIiIiIjEMCVtIiIiIiIiMUxJ24H1DToAkQKgz7kkOn3GpTDQ51wKg0L7OdecNhERERERkRimkTYREREREZEYpqRtP5xzFzjn5jrn5jnnHgg6HpFIcM6955xb6Zyble2+is65b51zf2ReVwgyRpH8cM4d6Zz73jn3u3PuN+dc98z79TmXhOCcK+mcm+qcm5n5GX8y8/7azrkpmcctnzjnSgQdq0h+OeeKOud+ds6NyPy60H7OlbTlwDlXFHgduBA4AbjGOXdCsFGJREQ/4IK97nsAGOO9rwuMyfxaJF7tAv7jvT8BOAPomvn3W59zSRTbgVbe+wbAqcAFzrkzgOeBXt77Y4G1QOfgQhSJmO7A7GxfF9rPuZK2nDUG5nnv53vvdwAfA5cFHJNIvnnvxwH/7HX3ZUD/zNv9gcsLMiaRSPLeL/PeT8+8vRH7Z18Dfc4lQXizKfPL4pkXD7QCBmfer8+4xD3nXE2gLfBO5teOQvw5V9KWsxrAomxfL868TyQRVfXeL8u8vRyoGmQwIpHinKsFNASmoM+5JJDMkrEZwErgW+BPYJ33flfmU3TcIongZeC/QEbm15UoxJ9zJW0iksVbO1m1lJW455wrA3wO3O2935D9MX3OJd5579O996cCNbHqoHrBRiQSWc65i4GV3vtpQccSK4oFHUCMWgIcme3rmpn3iSSiFc65at77Zc65atiZW5G45ZwrjiVsg7z3QzLv1udcEo73fp1z7nugKVDeOVcscxRCxy0S784CLnXOXQSUBMoBr1CIP+caacvZT0DdzA41JYCOwBcBxyQSLV8AN2XevgkYHmAsIvmSOefhXWC29/6lbA/pcy4JwTlXxTlXPvN2KaANNnfze6BD5tP0GZe45r1/0Htf03tfCzsOT/HeX0ch/pxrce39yMzsXwaKAu95758JNiKR/HPOfQS0ACoDK4DHgWHAp8BRwELgKu/93s1KROKCc64Z8CPwK7vnQTyEzWvT51zinnPuFKwBQ1Hs5Pun3vsezrk6WOO0isDPwPXe++3BRSoSGc65FsC93vuLC/PnXEmbiIiIiIhIDFN5pIiIiIiISAxT0iYiIiIiIhLDlLSJiIiIiIjEMCVtIiIiIiIiMUxJm4iIiIiISAxT0iYiIiIiIhLDlLSJiIiIiIjEMCVtIiIiIiIiMez/AYoTeJrDbjsEAAAAAElFTkSuQmCC\n",
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {
+ "needs_background": "light"
+ },
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "residuals = abs(np.array(valandtest_pred)-np.array(valandtest_orig))\n",
+ "residuals_df=pd.DataFrame()\n",
+ "residuals_df['residuals']=residuals\n",
+ "#residuals_df['ds']=prophet_dataset2['ds']\n",
+ "\n",
+ "f, ax = plt.subplots(figsize=(15,7))\n",
+ "ax.text(x=len(val_extract)/2 - 10,y=150, s='Validation Data', color='black', fontsize=15, fontweight='bold', alpha=0.8)\n",
+ "ax.text(x=len(val_extract)+10,y=150, s='Testing Data', color='black', fontsize=15, fontweight='bold', alpha=0.8)\n",
+ "ax.axvline(x=len(val_extract), c='green', lw=3, alpha=0.8, ls='--')\n",
+ "residuals_df.plot(kind='line', y='residuals', color='blue', label='Residual',figsize=(15, 6),ax=ax)\n",
+ "plt.title(\"Residuals\")"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "id": "inside-amber",
+ "metadata": {},
+ "source": [
+ "# Save the Predictions"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 28,
+ "id": "alert-recruitment",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "predictions_df=pd.DataFrame()\n",
+ "predictions_df['original'] = valandtest_orig\n",
+ "predictions_df['predictions'] = valandtest_pred\n",
+ "predictions_df.to_csv(Path(\"Predictions/Log.csv\"),index=False)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "id": "blocked-dakota",
+ "metadata": {},
+ "outputs": [],
+ "source": []
+ }
+ ],
+ "metadata": {
+ "kernelspec": {
+ "display_name": "Python 3",
+ "language": "python",
+ "name": "python3"
+ },
+ "language_info": {
+ "codemirror_mode": {
+ "name": "ipython",
+ "version": 3
+ },
+ "file_extension": ".py",
+ "mimetype": "text/x-python",
+ "name": "python",
+ "nbconvert_exporter": "python",
+ "pygments_lexer": "ipython3",
+ "version": "3.7.10"
+ },
+ "toc-autonumbering": true
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}
diff --git a/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/no_transformation.ipynb b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/no_transformation.ipynb
new file mode 100644
index 0000000000000000000000000000000000000000..5876bdd0fee7867f3a64918cdf24b655247ccc91
--- /dev/null
+++ b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/no_transformation.ipynb
@@ -0,0 +1,1752 @@
+{
+ "cells": [
+ {
+ "cell_type": "code",
+ "execution_count": 1,
+ "id": "unique-lightweight",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "%matplotlib inline\n",
+ "import mxnet as mx\n",
+ "from mxnet import gluon\n",
+ "import numpy as np\n",
+ "import pandas as pd\n",
+ "import matplotlib.pyplot as plt\n",
+ "import json\n",
+ "from gluonts.dataset.common import ListDataset\n",
+ "from datetime import datetime, timedelta\n",
+ "from gluonts.model.simple_feedforward import SimpleFeedForwardEstimator\n",
+ "from gluonts.model.deepar import DeepAREstimator\n",
+ "from gluonts.mx.trainer import Trainer\n",
+ "from gluonts.evaluation.backtest import make_evaluation_predictions\n",
+ "from gluonts.evaluation import Evaluator\n",
+ "from gluonts.dataset.util import to_pandas\n",
+ "from gluonts.dataset.field_names import FieldName\n",
+ "from gluonts.model.forecast import SampleForecast\n",
+ "from pandas import Timestamp\n",
+ "import itertools\n",
+ "from gluonts.model.estimator import GluonEstimator"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "id": "identical-holly",
+ "metadata": {},
+ "source": [
+ "# Data Loading"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 2,
+ "id": "lyric-netherlands",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "filename='EstimatedRemainingTimeContext'"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 3,
+ "id": "legislative-diesel",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "dataset= pd.read_csv(filename + \".csv\")\n",
+ "X= dataset[['name','time', 'countryCode', 'ipAddress', 'level',\n",
+ " 'producer']]\n",
+ "y= dataset.iloc[:,6]\n",
+ "#changing the names of the attributes\n",
+ "gluon_dataset= pd.DataFrame()\n",
+ "gluon_dataset['ds'] = pd.to_datetime(X[\"time\"])\n",
+ "gluon_dataset['y']=y\n",
+ "gluon_dataset['ds'] = gluon_dataset['ds'].dt.tz_convert(None)\n",
+ "gluon_dataset_copy=gluon_dataset.copy()"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 4,
+ "id": "similar-serbia",
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/html": [
+ "\n",
+ "\n",
+ "
\n",
+ " \n",
+ " \n",
+ " | \n",
+ " ds | \n",
+ " y | \n",
+ "
\n",
+ " \n",
+ " \n",
+ " \n",
+ " | 0 | \n",
+ " 2021-02-18 11:31:12.752 | \n",
+ " 4050.000000 | \n",
+ "
\n",
+ " \n",
+ " | 1 | \n",
+ " 2021-02-18 11:31:32.752 | \n",
+ " 4045.500000 | \n",
+ "
\n",
+ " \n",
+ " | 2 | \n",
+ " 2021-02-18 11:32:02.752 | \n",
+ " 4045.500000 | \n",
+ "
\n",
+ " \n",
+ " | 3 | \n",
+ " 2021-02-18 11:32:32.752 | \n",
+ " 4045.500000 | \n",
+ "
\n",
+ " \n",
+ " | 4 | \n",
+ " 2021-02-18 11:33:02.752 | \n",
+ " 4153.250000 | \n",
+ "
\n",
+ " \n",
+ " | ... | \n",
+ " ... | \n",
+ " ... | \n",
+ "
\n",
+ " \n",
+ " | 135 | \n",
+ " 2021-02-18 12:38:32.752 | \n",
+ " 110.500000 | \n",
+ "
\n",
+ " \n",
+ " | 136 | \n",
+ " 2021-02-18 12:39:02.752 | \n",
+ " 60.156250 | \n",
+ "
\n",
+ " \n",
+ " | 137 | \n",
+ " 2021-02-18 12:39:32.752 | \n",
+ " 19.140625 | \n",
+ "
\n",
+ " \n",
+ " | 138 | \n",
+ " 2021-02-18 12:40:02.752 | \n",
+ " 5.359375 | \n",
+ "
\n",
+ " \n",
+ " | 139 | \n",
+ " 2021-02-18 12:40:32.752 | \n",
+ " 0.765625 | \n",
+ "
\n",
+ " \n",
+ "
\n",
+ "
140 rows × 2 columns
\n",
+ "
"
+ ]
+ },
+ "metadata": {
+ "needs_background": "light"
+ },
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "train_entry = next(iter(train_ds))\n",
+ "train_entry.keys()\n",
+ "\n",
+ "validation_entry = next(iter(validation_ds))\n",
+ "validation_entry.keys()\n",
+ "\n",
+ "test_entry = next(iter(test_ds))\n",
+ "test_entry.keys()\n",
+ "\n",
+ "\n",
+ "test_series = to_pandas(test_entry)\n",
+ "train_series = to_pandas(train_entry)\n",
+ "validation_series = to_pandas(validation_entry)\n",
+ "\n",
+ "prediction_length = len(test_series) - len(validation_series)\n",
+ "context_length = [50,60]\n",
+ "\n",
+ "\n",
+ "list1=train_series.index\n",
+ "list2=train_entry['target']\n",
+ "\n",
+ "d = {'ds':list1,'y':list2}\n",
+ "df_train=pd.DataFrame(d, columns=['ds','y'])\n",
+ "\n",
+ "fig, ax = plt.subplots( figsize=(10, 7))\n",
+ "\n",
+ "# train_series.plot(ax=ax[0])\n",
+ "# ax[0].grid(which=\"both\")\n",
+ "# ax[0].legend([\"train series\"], loc=\"upper left\")\n",
+ "\n",
+ "test_series.plot(ax=ax)\n",
+ "plt.axvline(train_series.index[-1], color='r') # end of train dataset\n",
+ "plt.axvline(validation_series.index[-1], color='b') # end of validation dataset\n",
+ "ax.grid(which=\"both\")\n",
+ "ax.legend([\"test series\", \"end of train series\",\"end of validation series\"], loc=\"upper left\")\n",
+ "\n",
+ "plt.show()"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 9,
+ "id": "minus-brooks",
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "21"
+ ]
+ },
+ "execution_count": 9,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "prediction_length"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "id": "manual-cleveland",
+ "metadata": {},
+ "source": [
+ "# Hyper-parameter Tuning"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 10,
+ "id": "ethical-shoulder",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "batch_size = [75,100]\n",
+ "epochs = [5,10]\n",
+ "num_batches_per_epoch = [10,20]\n",
+ "learning_rate = [1e-3,1e-4]"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 11,
+ "id": "divine-seeking",
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "Total Possible Models 32\n"
+ ]
+ }
+ ],
+ "source": [
+ "from sklearn.model_selection import ParameterGrid\n",
+ "\n",
+ "param_grid = {'batch_size': batch_size,\n",
+ " 'epochs': epochs,\n",
+ " 'num_batches_per_epoch': num_batches_per_epoch,\n",
+ " 'learning_rate': learning_rate,\n",
+ " 'context_length': context_length\n",
+ " }\n",
+ "grid = ParameterGrid(param_grid)\n",
+ "cnt = 0\n",
+ "for p in grid:\n",
+ " cnt = cnt+1\n",
+ "\n",
+ "print('Total Possible Models',cnt)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 12,
+ "id": "phantom-algebra",
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stderr",
+ "output_type": "stream",
+ "text": [
+ " 0%| | 0/10 [00:00"
+ ]
+ },
+ "metadata": {
+ "needs_background": "light"
+ },
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "def plot_prob_forecasts(ts_entry, forecast_entry):\n",
+ " plot_length = len(ts_entry)\n",
+ " prediction_intervals = (50.0,98.0)\n",
+ " legend = [\"observations\", \"median prediction\"] + [f\"{k}% prediction interval\" for k in prediction_intervals][::-1]\n",
+ "\n",
+ " fig, ax = plt.subplots(1, 1, figsize=(10, 7))\n",
+ " ts_entry[-plot_length:].plot(ax=ax) # plot the time series\n",
+ " forecast_entry.plot(prediction_intervals=prediction_intervals, color='g')\n",
+ " plt.grid(which=\"both\")\n",
+ " plt.legend(legend, loc=\"upper left\")\n",
+ " plt.show()\n",
+ "plot_prob_forecasts(gluon_dataset['y'], forecast_entry)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 18,
+ "id": "atmospheric-thanks",
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stderr",
+ "output_type": "stream",
+ "text": [
+ "Running evaluation: 100%|██████████| 1/1 [00:00<00:00, 5.63it/s]\n"
+ ]
+ }
+ ],
+ "source": [
+ "evaluator = Evaluator(quantiles=[0.1, 0.5, 0.9])\n",
+ "agg_metrics, item_metrics = evaluator(iter(tss), iter([forecast_entry]), num_series=len(test_ds))\n"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 19,
+ "id": "southwest-undergraduate",
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "{'MSE': 354766.61904761905,\n",
+ " 'abs_error': 12105.1337890625,\n",
+ " 'abs_target_sum': 4846.515625,\n",
+ " 'abs_target_mean': 230.78645833333334,\n",
+ " 'seasonal_error': 86.95233050847457,\n",
+ " 'MASE': 6.629321364539261,\n",
+ " 'MAPE': 62.452435257275035,\n",
+ " 'sMAPE': 1.159284458125296,\n",
+ " 'OWA': nan,\n",
+ " 'MSIS': 208.114852388088,\n",
+ " 'QuantileLoss[0.1]': 18653.952172851565,\n",
+ " 'Coverage[0.1]': 1.0,\n",
+ " 'QuantileLoss[0.5]': 12105.134765625,\n",
+ " 'Coverage[0.5]': 1.0,\n",
+ " 'QuantileLoss[0.9]': 2775.030053710937,\n",
+ " 'Coverage[0.9]': 1.0,\n",
+ " 'RMSE': 595.6228832471256,\n",
+ " 'NRMSE': 2.580839827208777,\n",
+ " 'ND': 2.497698290008608,\n",
+ " 'wQuantileLoss[0.1]': 3.848940892014882,\n",
+ " 'wQuantileLoss[0.5]': 2.4976984915064624,\n",
+ " 'wQuantileLoss[0.9]': 0.572582504303994,\n",
+ " 'mean_absolute_QuantileLoss': 11178.038997395835,\n",
+ " 'mean_wQuantileLoss': 2.306407295941779,\n",
+ " 'MAE_Coverage': 0.5}"
+ ]
+ },
+ "execution_count": 19,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "agg_metrics"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "id": "extraordinary-arbitration",
+ "metadata": {},
+ "source": [
+ "# Residual"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 20,
+ "id": "affecting-generation",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "#get the mean for all the samples\n",
+ "f1 = forecast_entry_val.samples.mean(axis=0)\n",
+ "f2 = forecast_entry.samples.mean(axis=0)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 21,
+ "id": "institutional-blade",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "original_values=list(gluon_dataset['y'])\n",
+ "orig_test = original_values[:len(original_values)-prediction_length]\n",
+ "test_extract=original_values[-prediction_length:]\n",
+ "val_extract=orig_test[-prediction_length:]"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 22,
+ "id": "running-afternoon",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "valandtest_orig=list()\n",
+ "valandtest_pred=list()\n",
+ "for a1 in val_extract:\n",
+ " valandtest_orig.append(a1)\n",
+ "\n",
+ "for b1 in test_extract:\n",
+ " valandtest_orig.append(b1)\n",
+ " \n",
+ "for a2 in f1:\n",
+ " valandtest_pred.append(a2)\n",
+ "\n",
+ "for b2 in f2:\n",
+ " valandtest_pred.append(b2)\n"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 23,
+ "id": "every-venture",
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "Text(0.5, 1.0, 'Residuals')"
+ ]
+ },
+ "execution_count": 23,
+ "metadata": {},
+ "output_type": "execute_result"
+ },
+ {
+ "data": {
+ "image/png": "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\n",
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {
+ "needs_background": "light"
+ },
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "residuals = abs(np.array(valandtest_pred)-np.array(valandtest_orig))\n",
+ "residuals_df=pd.DataFrame()\n",
+ "residuals_df['residuals']=residuals\n",
+ "#residuals_df['ds']=prophet_dataset2['ds']\n",
+ "\n",
+ "f, ax = plt.subplots(figsize=(15,7))\n",
+ "ax.text(x=len(val_extract)/2 -10,y=400, s='Validation Data', color='black', fontsize=15, fontweight='bold', alpha=0.8)\n",
+ "ax.text(x=len(val_extract)+10,y=400, s='Testing Data', color='black', fontsize=15, fontweight='bold', alpha=0.8)\n",
+ "ax.axvline(x=len(val_extract), c='green', lw=3, alpha=0.8, ls='--')\n",
+ "residuals_df.plot(kind='line', y='residuals', color='blue', label='Residual',figsize=(15, 6),ax=ax)\n",
+ "plt.title(\"Residuals\")"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "id": "convinced-clause",
+ "metadata": {},
+ "source": [
+ "# Save the predictions"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 24,
+ "id": "antique-scoop",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "predictions_df=pd.DataFrame()\n",
+ "predictions_df['original'] = valandtest_orig\n",
+ "predictions_df['predictions'] = valandtest_pred\n",
+ "predictions_df.to_csv(Path(\"Predictions/NoTransformation.csv\"),index=False)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "id": "extraordinary-adaptation",
+ "metadata": {},
+ "outputs": [],
+ "source": []
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "id": "scenic-slope",
+ "metadata": {},
+ "outputs": [],
+ "source": []
+ }
+ ],
+ "metadata": {
+ "kernelspec": {
+ "display_name": "Python 3",
+ "language": "python",
+ "name": "python3"
+ },
+ "language_info": {
+ "codemirror_mode": {
+ "name": "ipython",
+ "version": 3
+ },
+ "file_extension": ".py",
+ "mimetype": "text/x-python",
+ "name": "python",
+ "nbconvert_exporter": "python",
+ "pygments_lexer": "ipython3",
+ "version": "3.7.10"
+ },
+ "toc-autonumbering": true
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}
diff --git a/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/squareroot.ipynb b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/squareroot.ipynb
new file mode 100644
index 0000000000000000000000000000000000000000..732122315fc0dbd2c89fc05bd1c910d29e5cce92
--- /dev/null
+++ b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/squareroot.ipynb
@@ -0,0 +1,2123 @@
+{
+ "cells": [
+ {
+ "cell_type": "code",
+ "execution_count": 1,
+ "id": "thorough-spine",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "%matplotlib inline\n",
+ "import mxnet as mx\n",
+ "from mxnet import gluon\n",
+ "import numpy as np\n",
+ "import pandas as pd\n",
+ "import matplotlib.pyplot as plt\n",
+ "import json\n",
+ "from gluonts.dataset.common import ListDataset\n",
+ "from datetime import datetime, timedelta\n",
+ "from gluonts.model.simple_feedforward import SimpleFeedForwardEstimator\n",
+ "from gluonts.model.deepar import DeepAREstimator\n",
+ "from gluonts.mx.trainer import Trainer\n",
+ "from gluonts.evaluation.backtest import make_evaluation_predictions\n",
+ "from gluonts.evaluation import Evaluator\n",
+ "from gluonts.dataset.util import to_pandas\n",
+ "from gluonts.dataset.field_names import FieldName\n",
+ "from gluonts.model.forecast import SampleForecast\n",
+ "from pandas import Timestamp\n",
+ "import itertools\n",
+ "from gluonts.model.estimator import GluonEstimator"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "id": "prescription-battle",
+ "metadata": {},
+ "source": [
+ "# Data Loading"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 2,
+ "id": "colored-geometry",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "filename='EstimatedRemainingTimeContext'"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 3,
+ "id": "infectious-omega",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "dataset= pd.read_csv(filename + \".csv\")\n",
+ "X= dataset[['name','time', 'countryCode', 'ipAddress', 'level',\n",
+ " 'producer']]\n",
+ "y= dataset.iloc[:,6]\n",
+ "#changing the names of the attributes\n",
+ "gluon_dataset= pd.DataFrame()\n",
+ "gluon_dataset['ds'] = pd.to_datetime(X[\"time\"])\n",
+ "gluon_dataset['y']=y\n",
+ "gluon_dataset['ds'] = gluon_dataset['ds'].dt.tz_convert(None)\n",
+ "gluon_dataset_copy=gluon_dataset.copy()"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 4,
+ "id": "patient-northern",
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/html": [
+ "\n",
+ "\n",
+ "
\n",
+ " \n",
+ " \n",
+ " | \n",
+ " ds | \n",
+ " y | \n",
+ "
\n",
+ " \n",
+ " \n",
+ " \n",
+ " | 0 | \n",
+ " 2021-02-18 11:31:12.752 | \n",
+ " 4050.000000 | \n",
+ "
\n",
+ " \n",
+ " | 1 | \n",
+ " 2021-02-18 11:31:32.752 | \n",
+ " 4045.500000 | \n",
+ "
\n",
+ " \n",
+ " | 2 | \n",
+ " 2021-02-18 11:32:02.752 | \n",
+ " 4045.500000 | \n",
+ "
\n",
+ " \n",
+ " | 3 | \n",
+ " 2021-02-18 11:32:32.752 | \n",
+ " 4045.500000 | \n",
+ "
\n",
+ " \n",
+ " | 4 | \n",
+ " 2021-02-18 11:33:02.752 | \n",
+ " 4153.250000 | \n",
+ "
\n",
+ " \n",
+ " | ... | \n",
+ " ... | \n",
+ " ... | \n",
+ "
\n",
+ " \n",
+ " | 135 | \n",
+ " 2021-02-18 12:38:32.752 | \n",
+ " 110.500000 | \n",
+ "
\n",
+ " \n",
+ " | 136 | \n",
+ " 2021-02-18 12:39:02.752 | \n",
+ " 60.156250 | \n",
+ "
\n",
+ " \n",
+ " | 137 | \n",
+ " 2021-02-18 12:39:32.752 | \n",
+ " 19.140625 | \n",
+ "
\n",
+ " \n",
+ " | 138 | \n",
+ " 2021-02-18 12:40:02.752 | \n",
+ " 5.359375 | \n",
+ "
\n",
+ " \n",
+ " | 139 | \n",
+ " 2021-02-18 12:40:32.752 | \n",
+ " 0.765625 | \n",
+ "
\n",
+ " \n",
+ "
\n",
+ "
140 rows × 2 columns
\n",
+ "
"
+ ]
+ },
+ "metadata": {
+ "needs_background": "light"
+ },
+ "output_type": "display_data"
+ },
+ {
+ "data": {
+ "image/png": "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\n",
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {
+ "needs_background": "light"
+ },
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "gluon_dataset_copy['y_orig']=gluon_dataset_copy['y']\n",
+ "transformed = square_root_transformation(gluon_dataset_copy['y'])\n",
+ "gluon_dataset_copy['y']=transformed\n",
+ "\n",
+ "#plot before transformation\n",
+ "pd.plotting.register_matplotlib_converters()\n",
+ "f, ax = plt.subplots(figsize=(15,7))\n",
+ "gluon_dataset_copy.plot(kind='line', x='ds', y='y_orig', color='blue', label='Train', ax=ax)\n",
+ "plt.title(filename)\n",
+ "plt.show()\n",
+ "\n",
+ "#plot after transformation\n",
+ "pd.plotting.register_matplotlib_converters()\n",
+ "f, ax = plt.subplots(figsize=(15,7))\n",
+ "gluon_dataset_copy.plot(kind='line', x='ds', y='y', color='blue', label='Train', ax=ax)\n",
+ "#test.plot(kind='line', x='ds', y='y', color='red', label='Test', ax=ax)\n",
+ "plt.title(filename + \"_transformed using Square Root Transformation\")\n",
+ "plt.show()"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 8,
+ "id": "alone-migration",
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/html": [
+ "\n",
+ "\n",
+ "
\n",
+ " \n",
+ " \n",
+ " | \n",
+ " ds | \n",
+ " y | \n",
+ " y_orig | \n",
+ "
\n",
+ " \n",
+ " \n",
+ " \n",
+ " | 0 | \n",
+ " 2021-02-18 11:31:12.752 | \n",
+ " 63.639610 | \n",
+ " 4050.000000 | \n",
+ "
\n",
+ " \n",
+ " | 1 | \n",
+ " 2021-02-18 11:32:12.752 | \n",
+ " 63.604245 | \n",
+ " 4045.500000 | \n",
+ "
\n",
+ " \n",
+ " | 2 | \n",
+ " 2021-02-18 11:33:12.752 | \n",
+ " 63.604245 | \n",
+ " 4045.500000 | \n",
+ "
\n",
+ " \n",
+ " | 3 | \n",
+ " 2021-02-18 11:34:12.752 | \n",
+ " 63.604245 | \n",
+ " 4045.500000 | \n",
+ "
\n",
+ " \n",
+ " | 4 | \n",
+ " 2021-02-18 11:35:12.752 | \n",
+ " 64.445714 | \n",
+ " 4153.250000 | \n",
+ "
\n",
+ " \n",
+ " | ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ "
\n",
+ " \n",
+ " | 135 | \n",
+ " 2021-02-18 13:46:12.752 | \n",
+ " 10.511898 | \n",
+ " 110.500000 | \n",
+ "
\n",
+ " \n",
+ " | 136 | \n",
+ " 2021-02-18 13:47:12.752 | \n",
+ " 7.756046 | \n",
+ " 60.156250 | \n",
+ "
\n",
+ " \n",
+ " | 137 | \n",
+ " 2021-02-18 13:48:12.752 | \n",
+ " 4.375000 | \n",
+ " 19.140625 | \n",
+ "
\n",
+ " \n",
+ " | 138 | \n",
+ " 2021-02-18 13:49:12.752 | \n",
+ " 2.315032 | \n",
+ " 5.359375 | \n",
+ "
\n",
+ " \n",
+ " | 139 | \n",
+ " 2021-02-18 13:50:12.752 | \n",
+ " 0.875000 | \n",
+ " 0.765625 | \n",
+ "
\n",
+ " \n",
+ "
\n",
+ "
140 rows × 3 columns
\n",
+ "
"
+ ]
+ },
+ "metadata": {
+ "needs_background": "light"
+ },
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "train_entry = next(iter(train_ds))\n",
+ "train_entry.keys()\n",
+ "\n",
+ "validation_entry = next(iter(validation_ds))\n",
+ "validation_entry.keys()\n",
+ "\n",
+ "test_entry = next(iter(test_ds))\n",
+ "test_entry.keys()\n",
+ "\n",
+ "test_orig_entry = next(iter(test_orig_ds))\n",
+ "test_orig_entry.keys()\n",
+ "\n",
+ "test_series = to_pandas(test_entry)\n",
+ "train_series = to_pandas(train_entry)\n",
+ "validation_series = to_pandas(validation_entry)\n",
+ "\n",
+ "prediction_length = len(test_series) - len(validation_series)\n",
+ "context_length = [50,60]\n",
+ "\n",
+ "\n",
+ "list1=train_series.index\n",
+ "list2=train_entry['target']\n",
+ "\n",
+ "d = {'ds':list1,'y':list2}\n",
+ "df_train=pd.DataFrame(d, columns=['ds','y'])\n",
+ "\n",
+ "fig, ax = plt.subplots( figsize=(10, 7))\n",
+ "\n",
+ "# train_series.plot(ax=ax[0])\n",
+ "# ax[0].grid(which=\"both\")\n",
+ "# ax[0].legend([\"train series\"], loc=\"upper left\")\n",
+ "\n",
+ "test_series.plot(ax=ax)\n",
+ "plt.axvline(train_series.index[-1], color='r') # end of train dataset\n",
+ "plt.axvline(validation_series.index[-1], color='b') # end of validation dataset\n",
+ "ax.grid(which=\"both\")\n",
+ "ax.legend([\"test series\", \"end of train series\",\"end of validation series\"], loc=\"upper left\")\n",
+ "\n",
+ "plt.show()"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 12,
+ "id": "minimal-central",
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "21"
+ ]
+ },
+ "execution_count": 12,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "prediction_length"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "id": "saved-magic",
+ "metadata": {},
+ "source": [
+ "# Hyper-parameter Tuning"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 13,
+ "id": "listed-stack",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "batch_size = [75,100]\n",
+ "epochs = [5,10]\n",
+ "num_batches_per_epoch = [10,20]\n",
+ "learning_rate = [1e-3,1e-4]"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 14,
+ "id": "rough-hands",
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "Total Possible Models 32\n"
+ ]
+ }
+ ],
+ "source": [
+ "from sklearn.model_selection import ParameterGrid\n",
+ "\n",
+ "param_grid = {#'num_hidden_dimensions':num_hidden_dimensions,\n",
+ " 'batch_size': batch_size,\n",
+ " 'epochs': epochs,\n",
+ " 'num_batches_per_epoch': num_batches_per_epoch,\n",
+ " 'learning_rate': learning_rate,\n",
+ " 'context_length': context_length\n",
+ " }\n",
+ "grid = ParameterGrid(param_grid)\n",
+ "cnt = 0\n",
+ "for p in grid:\n",
+ " cnt = cnt+1\n",
+ "\n",
+ "print('Total Possible Models',cnt)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 15,
+ "id": "twelve-desperate",
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stderr",
+ "output_type": "stream",
+ "text": [
+ " 0%| | 0/10 [00:00"
+ ]
+ },
+ "metadata": {
+ "needs_background": "light"
+ },
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "def plot_prob_forecasts(ts_entry, forecast_entry):\n",
+ " plot_length = len(ts_entry)\n",
+ " prediction_intervals = (50.0,98.0)\n",
+ " legend = [\"observations\", \"median prediction\"] + [f\"{k}% prediction interval\" for k in prediction_intervals][::-1]\n",
+ "\n",
+ " fig, ax = plt.subplots(1, 1, figsize=(10, 7))\n",
+ " ts_entry[-plot_length:].plot(ax=ax) # plot the time series\n",
+ " forecast_entry.plot(prediction_intervals=prediction_intervals, color='g')\n",
+ " plt.grid(which=\"both\")\n",
+ " plt.legend(legend, loc=\"upper left\")\n",
+ " plt.show()\n",
+ "plot_prob_forecasts(gluon_dataset['y'], forecast_entry_inv)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 22,
+ "id": "developmental-reset",
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stderr",
+ "output_type": "stream",
+ "text": [
+ "Running evaluation: 100%|██████████| 1/1 [00:00<00:00, 5.33it/s]\n"
+ ]
+ }
+ ],
+ "source": [
+ "evaluator = Evaluator(quantiles=[0.1, 0.5, 0.9])\n",
+ "agg_metrics, item_metrics = evaluator(iter(tss_inv), iter([forecast_entry_inv]), num_series=len(test_ds))\n"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 23,
+ "id": "noted-emergency",
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "{'MSE': 2491714.285714286,\n",
+ " 'abs_error': 31313.115234375,\n",
+ " 'abs_target_sum': 4846.515625,\n",
+ " 'abs_target_mean': 230.78645833333334,\n",
+ " 'seasonal_error': 86.95233050847457,\n",
+ " 'MASE': 17.148484885072776,\n",
+ " 'MAPE': 165.68347366521184,\n",
+ " 'sMAPE': 1.4677628931523303,\n",
+ " 'OWA': nan,\n",
+ " 'MSIS': 524.5641392017257,\n",
+ " 'QuantileLoss[0.1]': 49122.942187500004,\n",
+ " 'Coverage[0.1]': 1.0,\n",
+ " 'QuantileLoss[0.5]': 31313.116149902344,\n",
+ " 'Coverage[0.5]': 1.0,\n",
+ " 'QuantileLoss[0.9]': 7017.614105224608,\n",
+ " 'Coverage[0.9]': 1.0,\n",
+ " 'RMSE': 1578.516482560219,\n",
+ " 'NRMSE': 6.839727486438176,\n",
+ " 'ND': 6.460954148760224,\n",
+ " 'wQuantileLoss[0.1]': 10.13572347401645,\n",
+ " 'wQuantileLoss[0.5]': 6.460954337664463,\n",
+ " 'wQuantileLoss[0.9]': 1.4479710060203526,\n",
+ " 'mean_absolute_QuantileLoss': 29151.22414754232,\n",
+ " 'mean_wQuantileLoss': 6.014882939233755,\n",
+ " 'MAE_Coverage': 0.5}"
+ ]
+ },
+ "execution_count": 23,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "agg_metrics"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "id": "integrated-advancement",
+ "metadata": {},
+ "source": [
+ "# Residual"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 24,
+ "id": "beneficial-situation",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "#get the mean for all the samples\n",
+ "f1 = forecast_entry_val_inv.samples.mean(axis=0)\n",
+ "f2 = forecast_entry_inv.samples.mean(axis=0)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 25,
+ "id": "killing-guatemala",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "original_values=list(gluon_dataset['y'])\n",
+ "orig_test = original_values[:len(original_values)-prediction_length]\n",
+ "test_extract=original_values[-prediction_length:]\n",
+ "val_extract=orig_test[-prediction_length:]"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 26,
+ "id": "acquired-scientist",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "valandtest_orig=list()\n",
+ "valandtest_pred=list()\n",
+ "for a1 in val_extract:\n",
+ " valandtest_orig.append(a1)\n",
+ "\n",
+ "for b1 in test_extract:\n",
+ " valandtest_orig.append(b1)\n",
+ " \n",
+ "for a2 in f1:\n",
+ " valandtest_pred.append(a2)\n",
+ "\n",
+ "for b2 in f2:\n",
+ " valandtest_pred.append(b2)\n"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 27,
+ "id": "meaning-seeker",
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "Text(0.5, 1.0, 'Residuals')"
+ ]
+ },
+ "execution_count": 27,
+ "metadata": {},
+ "output_type": "execute_result"
+ },
+ {
+ "data": {
+ "image/png": "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\n",
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {
+ "needs_background": "light"
+ },
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "residuals = abs(np.array(valandtest_pred)-np.array(valandtest_orig))\n",
+ "residuals_df=pd.DataFrame()\n",
+ "residuals_df['residuals']=residuals\n",
+ "#residuals_df['ds']=prophet_dataset2['ds']\n",
+ "\n",
+ "f, ax = plt.subplots(figsize=(15,7))\n",
+ "ax.text(x=len(val_extract)/2 -10,y=400, s='Validation Data', color='black', fontsize=15, fontweight='bold', alpha=0.8)\n",
+ "ax.text(x=len(val_extract)+10,y=400, s='Testing Data', color='black', fontsize=15, fontweight='bold', alpha=0.8)\n",
+ "ax.axvline(x=len(val_extract), c='green', lw=3, alpha=0.8, ls='--')\n",
+ "residuals_df.plot(kind='line', y='residuals', color='blue', label='Residual',figsize=(15, 6),ax=ax)\n",
+ "plt.title(\"Residuals\")"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "id": "empirical-bristol",
+ "metadata": {},
+ "source": [
+ "# Save the Predictions"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 28,
+ "id": "awful-static",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "predictions_df=pd.DataFrame()\n",
+ "predictions_df['original'] = valandtest_orig\n",
+ "predictions_df['predictions'] = valandtest_pred\n",
+ "predictions_df.to_csv(Path(\"Predictions/SquareRoot.csv\"),index=False)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "id": "improving-cowboy",
+ "metadata": {},
+ "outputs": [],
+ "source": []
+ }
+ ],
+ "metadata": {
+ "kernelspec": {
+ "display_name": "Python 3",
+ "language": "python",
+ "name": "python3"
+ },
+ "language_info": {
+ "codemirror_mode": {
+ "name": "ipython",
+ "version": 3
+ },
+ "file_extension": ".py",
+ "mimetype": "text/x-python",
+ "name": "python",
+ "nbconvert_exporter": "python",
+ "pygments_lexer": "ipython3",
+ "version": "3.7.10"
+ },
+ "toc-autonumbering": true
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}
diff --git a/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/standardization.ipynb b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/standardization.ipynb
new file mode 100644
index 0000000000000000000000000000000000000000..1bcb5b37204562f66afdaa4a255b5ebad86c6d6c
--- /dev/null
+++ b/time-series-experiments/gluonts_activeeon/benchmark/genome/deployment-reconfiguration-range-1-to-10_ 2021-02-18 to 2021-02-18/gluon_deepar/notebooks/estimatedremainingtimecontext/standardization.ipynb
@@ -0,0 +1,1947 @@
+{
+ "cells": [
+ {
+ "cell_type": "code",
+ "execution_count": 1,
+ "id": "split-argentina",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "%matplotlib inline\n",
+ "import mxnet as mx\n",
+ "from mxnet import gluon\n",
+ "import numpy as np\n",
+ "import pandas as pd\n",
+ "import matplotlib.pyplot as plt\n",
+ "import json\n",
+ "from gluonts.dataset.common import ListDataset\n",
+ "from datetime import datetime, timedelta\n",
+ "from gluonts.model.simple_feedforward import SimpleFeedForwardEstimator\n",
+ "from gluonts.model.deepar import DeepAREstimator\n",
+ "from gluonts.mx.trainer import Trainer\n",
+ "from gluonts.evaluation.backtest import make_evaluation_predictions\n",
+ "from gluonts.evaluation import Evaluator\n",
+ "from gluonts.dataset.util import to_pandas\n",
+ "from gluonts.dataset.field_names import FieldName\n",
+ "from gluonts.model.forecast import SampleForecast\n",
+ "from pandas import Timestamp\n",
+ "import itertools\n",
+ "from gluonts.model.estimator import GluonEstimator\n",
+ "from numpy import array"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "id": "stock-excess",
+ "metadata": {},
+ "source": [
+ "# Data Loading"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 2,
+ "id": "suitable-teaching",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "filename='EstimatedRemainingTimeContext'"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 3,
+ "id": "mature-mumbai",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "dataset= pd.read_csv(filename + \".csv\")\n",
+ "X= dataset[['name','time', 'countryCode', 'ipAddress', 'level',\n",
+ " 'producer']]\n",
+ "y= dataset.iloc[:,6]\n",
+ "#changing the names of the attributes\n",
+ "gluon_dataset= pd.DataFrame()\n",
+ "gluon_dataset['ds'] = pd.to_datetime(X[\"time\"])\n",
+ "gluon_dataset['y']=y\n",
+ "gluon_dataset['ds'] = gluon_dataset['ds'].dt.tz_convert(None)\n",
+ "gluon_dataset_copy=gluon_dataset.copy()"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 4,
+ "id": "limited-arkansas",
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/html": [
+ "\n",
+ "\n",
+ "
\n",
+ " \n",
+ " \n",
+ " | \n",
+ " ds | \n",
+ " y | \n",
+ "
\n",
+ " \n",
+ " \n",
+ " \n",
+ " | 0 | \n",
+ " 2021-02-18 11:31:12.752 | \n",
+ " 4050.000000 | \n",
+ "
\n",
+ " \n",
+ " | 1 | \n",
+ " 2021-02-18 11:31:32.752 | \n",
+ " 4045.500000 | \n",
+ "
\n",
+ " \n",
+ " | 2 | \n",
+ " 2021-02-18 11:32:02.752 | \n",
+ " 4045.500000 | \n",
+ "
\n",
+ " \n",
+ " | 3 | \n",
+ " 2021-02-18 11:32:32.752 | \n",
+ " 4045.500000 | \n",
+ "
\n",
+ " \n",
+ " | 4 | \n",
+ " 2021-02-18 11:33:02.752 | \n",
+ " 4153.250000 | \n",
+ "
\n",
+ " \n",
+ " | ... | \n",
+ " ... | \n",
+ " ... | \n",
+ "
\n",
+ " \n",
+ " | 135 | \n",
+ " 2021-02-18 12:38:32.752 | \n",
+ " 110.500000 | \n",
+ "
\n",
+ " \n",
+ " | 136 | \n",
+ " 2021-02-18 12:39:02.752 | \n",
+ " 60.156250 | \n",
+ "
\n",
+ " \n",
+ " | 137 | \n",
+ " 2021-02-18 12:39:32.752 | \n",
+ " 19.140625 | \n",
+ "
\n",
+ " \n",
+ " | 138 | \n",
+ " 2021-02-18 12:40:02.752 | \n",
+ " 5.359375 | \n",
+ "
\n",
+ " \n",
+ " | 139 | \n",
+ " 2021-02-18 12:40:32.752 | \n",
+ " 0.765625 | \n",
+ "
\n",
+ " \n",
+ "
\n",
+ "
140 rows × 2 columns
\n",
+ "
"
+ ]
+ },
+ "metadata": {
+ "needs_background": "light"
+ },
+ "output_type": "display_data"
+ },
+ {
+ "data": {
+ "image/png": "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\n",
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {
+ "needs_background": "light"
+ },
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "from sklearn.preprocessing import StandardScaler\n",
+ "transformer = StandardScaler()\n",
+ "gluon_dataset_copy['y_orig']=gluon_dataset_copy['y']\n",
+ "data = array(gluon_dataset_copy['y']).reshape(len(gluon_dataset_copy['y']), 1)\n",
+ "transformer.fit(data)\n",
+ "transformed = transformer.transform(data)\n",
+ "gluon_dataset_copy['y']=transformed\n",
+ "\n",
+ "\n",
+ "#transformed = cube_root_transformation(gluon_dataset_copy['y'])\n",
+ "#gluon_dataset_copy['y']=transformed\n",
+ "\n",
+ "#plot before transformation\n",
+ "pd.plotting.register_matplotlib_converters()\n",
+ "f, ax = plt.subplots(figsize=(15,7))\n",
+ "gluon_dataset_copy.plot(kind='line', x='ds', y='y_orig', color='blue', label='Train', ax=ax)\n",
+ "plt.title(filename)\n",
+ "plt.show()\n",
+ "\n",
+ "#plot after transformation\n",
+ "pd.plotting.register_matplotlib_converters()\n",
+ "f, ax = plt.subplots(figsize=(15,7))\n",
+ "gluon_dataset_copy.plot(kind='line', x='ds', y='y', color='blue', label='Train', ax=ax)\n",
+ "#test.plot(kind='line', x='ds', y='y', color='red', label='Test', ax=ax)\n",
+ "plt.title(filename + \"_transformed using Square Root Transformation\")\n",
+ "plt.show()"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 7,
+ "id": "advisory-craps",
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/html": [
+ "\n",
+ "\n",
+ "
\n",
+ " \n",
+ " \n",
+ " | \n",
+ " ds | \n",
+ " y | \n",
+ " y_orig | \n",
+ "
\n",
+ " \n",
+ " \n",
+ " \n",
+ " | 0 | \n",
+ " 2021-02-18 11:31:12.752 | \n",
+ " 1.906777 | \n",
+ " 4050.000000 | \n",
+ "
\n",
+ " \n",
+ " | 1 | \n",
+ " 2021-02-18 11:32:12.752 | \n",
+ " 1.903170 | \n",
+ " 4045.500000 | \n",
+ "
\n",
+ " \n",
+ " | 2 | \n",
+ " 2021-02-18 11:33:12.752 | \n",
+ " 1.903170 | \n",
+ " 4045.500000 | \n",
+ "
\n",
+ " \n",
+ " | 3 | \n",
+ " 2021-02-18 11:34:12.752 | \n",
+ " 1.903170 | \n",
+ " 4045.500000 | \n",
+ "
\n",
+ " \n",
+ " | 4 | \n",
+ " 2021-02-18 11:35:12.752 | \n",
+ " 1.989524 | \n",
+ " 4153.250000 | \n",
+ "
\n",
+ " \n",
+ " | ... | \n",
+ " ... | \n",
+ " ... | \n",
+ " ... | \n",
+ "
\n",
+ " \n",
+ " | 135 | \n",
+ " 2021-02-18 13:46:12.752 | \n",
+ " -1.250452 | \n",
+ " 110.500000 | \n",
+ "
\n",
+ " \n",
+ " | 136 | \n",
+ " 2021-02-18 13:47:12.752 | \n",
+ " -1.290799 | \n",
+ " 60.156250 | \n",
+ "
\n",
+ " \n",
+ " | 137 | \n",
+ " 2021-02-18 13:48:12.752 | \n",
+ " -1.323670 | \n",
+ " 19.140625 | \n",
+ "
\n",
+ " \n",
+ " | 138 | \n",
+ " 2021-02-18 13:49:12.752 | \n",
+ " -1.334715 | \n",
+ " 5.359375 | \n",
+ "
\n",
+ " \n",
+ " | 139 | \n",
+ " 2021-02-18 13:50:12.752 | \n",
+ " -1.338396 | \n",
+ " 0.765625 | \n",
+ "
\n",
+ " \n",
+ "
\n",
+ "
140 rows × 3 columns
\n",
+ "