Learning under concept drift with follow the regularized leader and adaptive decaying proximal

作者：

Highlights：

• Propose a new adaptive learning algorithm to address the problem of concept drift.

• Use a decaying factor to discount previous learning examples.

• Use a concept drift detector to reset the learning process upon major concept drift.

• The proposed algorithm was theoretically proved to have sublinear regret bound.

摘要

•Propose a new adaptive learning algorithm to address the problem of concept drift.•Use a decaying factor to discount previous learning examples.•Use a concept drift detector to reset the learning process upon major concept drift.•The proposed algorithm was theoretically proved to have sublinear regret bound.

论文关键词：Concept drift,Decaying rate,Drift detector,Online learning,Follow the regularized leader

论文评审过程：Received 29 June 2017, Revised 20 November 2017, Accepted 21 November 2017, Available online 23 November 2017, Version of Record 22 December 2017.

论文官网地址：https://doi.org/10.1016/j.eswa.2017.11.042