I’m using the River ML library and working with decision tree models. I’d like to print or extract the decision rules from the trained trees. However, it seems that the library doesn’t currently offer a built-in method to do this. Does anyone know a way to print or extract the decision rules manually from a tree model in River? pipeline = compose.Pipeline( self.extract_features(event=event), OrdinalEncoder(), LASTClassifier(max_depth=5, track_error=True, remove_poor_attrs=True) ) metric = F1() self.state.update({'pipeline': pipeline, 'metric': metric}) label = 1 if classification_label else 0 # Retrieve current pipeline components pipeline = self.state.value()['pipeline'] model = pipeline['LASTClassifier'] metric = self.state.value()['metric'] # Make a prediction and update metric predicted_label = model.predict_one(self.state.value()['pipeline']['dict']) metric.update(y_true=label, y_pred=predicted_label) # Update the model with the new event model.learn_one(self.state.value()['pipeline']['dict'], label) # Save updated state self.state.update({'pipeline': pipeline, 'metric': metric}) Continue reading...
