Evolving rule-based classifiers with genetic programming on GPUs for drifting data streams

作者：

Highlights：

• Grammar-guided genetic programming rule-based classifier for drifting data streams.

• Online induction of highly accurate and interpretable rules.

• Fast adaptation to any type of concept drift.

• Mechanisms for rule diversification and adaptation.

• Efficient implementation on GPUs suitable for high-speed data streams.

摘要

•Grammar-guided genetic programming rule-based classifier for drifting data streams.•Online induction of highly accurate and interpretable rules.•Fast adaptation to any type of concept drift.•Mechanisms for rule diversification and adaptation.•Efficient implementation on GPUs suitable for high-speed data streams.

论文关键词：Machine learning,Data streams,Concept drift,Genetic programming,Rule-based classification,GPU,High-performance data mining

论文评审过程：Received 3 October 2017, Revised 31 July 2018, Accepted 21 October 2018, Available online 22 October 2018, Version of Record 27 October 2018.

论文官网地址：https://doi.org/10.1016/j.patcog.2018.10.024