On learning guarantees to unsupervised concept drift detection on data streams

作者：

Highlights：

• We present an approach to detect concept drifts on data streams.

• Our approach provides theoretical learning guarantees.

• McDiarmid’s inequality is employed to formalize model divergences.

摘要

•We present an approach to detect concept drifts on data streams.•Our approach provides theoretical learning guarantees.•McDiarmid’s inequality is employed to formalize model divergences.

论文关键词：Data streams,Concept drift,Algorithmic stability,McDiarmid’s inequality

论文评审过程：Received 21 June 2018, Revised 31 August 2018, Accepted 31 August 2018, Available online 21 September 2018, Version of Record 27 September 2018.

论文官网地址：https://doi.org/10.1016/j.eswa.2018.08.054