Reduce variability of sample dataset crossfold experiments

Created by: vcaballero-salle

We should try to reduce variability between samples during crossfold experiments. In thi way, we should be able to make better comparisons.

We identify two items to reduce variability:

• Do not vary sample dataset in the same split
• Do not vary sample dataset for multiple splits (WE SHOULD NOT DO THIS)

Then, the variability among splits only comes from the manner in which we clusterise prosumers, not from taking a random saple of a given size.