Boosting decision stumps for dynamic feature selection on data streams

作者：

Highlights：

• A novel method for dynamic feature selection for data streams is proposed.

• The proposal is evaluated on both real-world and synthetic scenarios.

• The method improves the accuracy rates of different types of data stream learners.

• Evaluation metrics for feature selection are proposed for streaming scenarios.

摘要

•A novel method for dynamic feature selection for data streams is proposed.•The proposal is evaluated on both real-world and synthetic scenarios.•The method improves the accuracy rates of different types of data stream learners.•Evaluation metrics for feature selection are proposed for streaming scenarios.

论文关键词：Data stream mining,Feature selection,Concept drift,Feature drift

论文评审过程：Received 26 June 2018, Revised 7 January 2019, Accepted 8 February 2019, Available online 13 February 2019, Version of Record 19 February 2019.

论文官网地址：https://doi.org/10.1016/j.is.2019.02.003