0

I used XGBClassifier to fit the dataset and got feature_score using codes below:

feature_score = clf.get_booster().get_score()

The fscores i got all greater than 1 but how can that possible?

bulala
  • 45
  • 1
  • 7

1 Answers1

0

It is not F1 score that XGBoost is using to assess feature importances. F1 score conveys the balance between precision and recall, and varies between [0, 1]. XGBoost assess the feature importances with a feature score, which tells the number of occurence of features in the ensemble, and can vary between [0, x], where x is the number of trees in your ensemble.

Adam Wilson
  • 138
  • 2
  • 6