OSFSMI: Online stream feature selection method based on mutual information

Maryam Rahmaninia, Parham Moradi

Abstract: Highlights•Two novel online streaming feature selection methods called OSFSMI and OSFSMI-k are proposed.•The proposed methods used mutual information to eliminate redundant and irrelevant features.•An incremental method is used to compute correlation of features in an online manner.•The proposed methods were compared with online and offline feature selection methods.•The reported results reveal that our methods are stable and scalable on high dimensional streaming data.