Add Khiops parameters to predictor constructors

[popescu-v]

Estimators of sklearn module (KhiopsClassifier, KhiopsRegressor and KhiopsEncoder) do not support all parameters that are available in the train_ functions of core module. The objective is therefore to extend their constructors by introducing the missing parameters.

1. For KhiopsClassifier and KhiopsRegressor, the following parameters are going to be added:

• max_evaluated_variables : int, default 0 Maximum number of variables to be evaluated in the SNB predictor training. If equal to 0 it evaluates all informative variables.
• max_selected_variables : int, default 0 Maximum number of variables to be selected in the SNB predictor. If equal to 0 it selects all the variables kept in the training.
• construction_rules : list of str, optional Allowed rules for the automatic variable construction. If not set it uses all possible rules.
• specific_pairs : list of tuple, optional User-specified pairs as a list of 2-tuples of variable names. If a given tuple contains only one non-empty string generated within the maximum limit max_pairs.
• all_possible_pairs : bool, default True If True tries to create all possible pairs within the limit max_pairs. The pairs and variables given in specific_pairs have priority.
• group_target_value : bool, default False Allows grouping of the target variable values in classification. It can substantially increase the training time.

2. For KhiopsEncoder, the following parameters are going to be added:
   • construction_rules : list of str, optional Allowed rules for the automatic variable construction. If not set it uses all possible rules.
   • specific_pairs : list of tuple, optional User-specified pairs as a list of 2-tuples of variable names. If a given tuple contains only one non-empty string generated within the maximum limit max_pairs.
   • all_possible_pairs : bool, default True If True tries to create all possible pairs within the limit max_pairs. The pairs and variables given in specific_pairs have priority.
   • group_target_value : bool, default False Allows grouping of the target variable values in classification. It can substantially increase the training time.
   • informative_variables_only : bool, default True If True keeps only informative variables.
   • pairs_recoding_method : str Type of recoding for bivariate variables. Types available:
     • "part Id" (default): An id for the interval/group
     • "part label": A label for the interval/group
     • "0-1 binarization": A 0's and 1's coding the interval/group id
     • "conditional info": Conditional information of the interval/group
     • "none": Keeps the variable as-is

[popescu-v]

• only few parameters are shown to the user compared to core.

[popescu-v]

• only few parameters are shown to the user compared to core.