Modéliser les consommations électriques de sites B2B ----------My solution to the challenge Datascience.net:

• pandas (pip install -U pandas)
• numpy (pip install -U numpy)
• sklearn (pip install -U sklearn)
• (matplotlib)

To test and run the implementation, use the driver `main`:

./main --verbose --use_cache_data --use_cache_trainingset --test --plot --compute_reality

A file named str(time.time())+'_ML.csv' will be created in ./results/ if istest == False. Else, the training is not done on the last ten days, and a comparison with the true value is done. For testing purpose, running:

%run main

in IPython, and play around with matplotlib is the best solution.

• nothing