How does feature_importances_ work in sklearn.tree?

Question

Python

I understand that feature_importances_ shows the importance of features, but for some reason it always gives me the type [0. 0. 0. 1.]. But how to understand which is the less significant parameter? And in general, what does it refer to?

Answer 1

feature_importances_returns a vector of feature "importances". The index of the element in this vector corresponds to the index of the feature in the data. The value of an element reflects the "importance" of the feature relative to the others: the larger the value, the greater the importance.