Feature ranking for enhancing boosting-based multi-label text categorization

作者：

Highlights：

• Several feature ranking methods are evaluated for improving RFBoost.

• An efficient version of RFBoost namely “RFBoost1” is proposed.

• BoWT representation model is used for representing text documents.

• Promising results are reported and RFBoost achieved the best performance.

• RFBoost1 proved efficient and it is a good alternative to AdaBoost.MH.

摘要

•Several feature ranking methods are evaluated for improving RFBoost.•An efficient version of RFBoost namely “RFBoost1” is proposed.•BoWT representation model is used for representing text documents.•Promising results are reported and RFBoost achieved the best performance.•RFBoost1 proved efficient and it is a good alternative to AdaBoost.MH.

论文关键词：RFBoost,Boosting,Multi-label learning,Text categorization,Feature ranking

论文评审过程：Received 24 January 2018, Revised 9 July 2018, Accepted 10 July 2018, Available online 21 July 2018, Version of Record 21 July 2018.

论文官网地址：https://doi.org/10.1016/j.eswa.2018.07.024