lightgbm 的 importance_type 'split' 和 xgboost 的 importance_type 'weight` 一样吗?

Is the importance_type 'split' of lightgbm the same as the importance_type 'weight` in xgboost?

lightgbm的importance_type'split'和xgboost的importance_type'weight'一样吗?

也就是说,下面是一样的吗?

booster.feature_importance(importance_type = 'split') # for lightgbm 

get_fscore(importance_type='weight') # for xgboost

尽管措辞略有不同,但它们确实是一样的。

来自 LightGBM docs:

If "split", result contains numbers of times the feature is used in a model.

来自 XGBoost docs:

'weight’: the number of times a feature is used to split the data across all trees.

这些重要性类型分别是两个包中的默认选择并非巧合。