Corrected commands to use data points from the beginning of the timeseries rather than clipping them, and using data from the test part of the set.
Receive configuration file as a command-line parameter for the benchmarking of the method
Simplified the csv.writer call to only require the output file as a parameter
Improved the generation of benchmarking output files to include both the predicted and observed values
Addition of the benchmark datasets to the 'datasets' folder