Leave K out split train validation, leaves some data out for validation even when the flag use_validation_set is False.

[pm390]

The interested file is this one:
RecSys_Course_AT_PoliMi/Data_manager/split_functions/split_train_validation_leave_k_out.py
lines 88 and 89.

Even if we set use_validation_set=False the validation data is held out (and used nowhere):

user_interaction_items_train = user_interaction_items[k_out*2:]
user_interaction_data_train = user_interaction_data[k_out*2:]

A possible change would be:

split_point= k_out*2 if use_validation_set else k_out
user_interaction_items_train = user_interaction_items[split_point:]
user_interaction_data_train = user_interaction_data[split_point:]