return local model

[oathaha]

I think it would be beneficial for user if local model of explain() function is included in returned dictionary.

The current code looks like this

rule_obj = {'synthetic_data': synthetic_instances,
                    'synthetic_predictions': synthetic_predictions,
                    'X_explain': X_explain,
                    'y_explain': y_explain,
                    'indep': self.indep,
                    'dep': self.dep,
                    'top_k_positive_rules': top_k_positive_rules,
                    'top_k_negative_rules': top_k_negative_rules}

but I think it should looks like this

rule_obj = {'synthetic_data': synthetic_instances,
                    'synthetic_predictions': synthetic_predictions,
                    'X_explain': X_explain,
                    'y_explain': y_explain,
                    'indep': self.indep,
                    'dep': self.dep,
                    'local_model': local_rulefit_model,
                    'top_k_positive_rules': top_k_positive_rules,
                    'top_k_negative_rules': top_k_negative_rules}

[MichaelFu1998-create]

Hi, it sounds great, could you provide the potential use case as an example so that I can understand your idea more clearly.

[oathaha]

Here is my idea local_model = rule_obj ['local_model'] # get local model from rule_obj

other code for experiment

In my opinion, some researchers would like to save local model for later use (like for evaluation). So it will be useful for them if they can obtain local model.

[MichaelFu1998-create]

Agree. Will update the code to implement this. Thanks!