search

scikit-learn-contrib / boruta_py

Python implementations of the Boruta all-relevant feature selection method.
BSD 3-Clause "New" or "Revised" License
1.46k stars 252 forks source link

donot suport lgb/xgb? #52

Closed yuanjie-ai closed 5 years ago

yuanjie-ai commented 5 years ago

donot suport lgb/xgb?

hatunina commented 5 years ago

Hi @Jie-Yuan , I also had the same question. I solved it by using lgb scikit-learn api . Because Boruta uses "featureimportances" internally.

However, it may not be wrong because it is not used as much yet. Sorry at that time.

marcospiau commented 5 years ago

I managed to get lgb working by following the instructions on the comments section of this kernel on Kaggle (https://www.kaggle.com/ogrellier/noise-analysis-of-porto-seguro-s-features). You need to change the line 282 of boruta_py.py to: self.estimator.set_params(random_state=self.random_state.get_state()[2])