Hi, I'm trying to understand how the train_or_test parameter is used.
I can see from the code that when it is set to 'test' the data (X_boruta) is split into train/test sets and the model fit on the training set (x_boruta_train). However, I don't understand how the importance is being calculated just on the test set.
The explain() function is calculating the shap values using the entire data (X_boruta), and not the test set (X_boruta_test):
self.shap_values = np.array(explainer.shap_values(self.X_boruta))
Am I missing something?
Hi, I'm trying to understand how the train_or_test parameter is used. I can see from the code that when it is set to 'test' the data (X_boruta) is split into train/test sets and the model fit on the training set (x_boruta_train). However, I don't understand how the importance is being calculated just on the test set. The explain() function is calculating the shap values using the entire data (X_boruta), and not the test set (X_boruta_test): self.shap_values = np.array(explainer.shap_values(self.X_boruta)) Am I missing something?