Refinement and selection heuristics in subgroup discovery and classification rule learning

作者：

Highlights：

• New double beam algorithms for subgroup discovery (SD) and classification rules (RL).

• Algorithms can use different heuristics for rule refinement and rule selection.

• Variants of new SD algorithm give more interesting rules than state-of-the-art.

• RL algorithm gives rules with comparable accuracy with state-of-the-art algorithms.

• Inverted heuristics in rule refinement produce rules with better coverage.

摘要

•New double beam algorithms for subgroup discovery (SD) and classification rules (RL).•Algorithms can use different heuristics for rule refinement and rule selection.•Variants of new SD algorithm give more interesting rules than state-of-the-art.•RL algorithm gives rules with comparable accuracy with state-of-the-art algorithms.•Inverted heuristics in rule refinement produce rules with better coverage.

论文关键词：Rule learning,Subgroup discovery,Inverted heuristics

论文评审过程：Received 7 June 2016, Revised 17 March 2017, Accepted 18 March 2017, Available online 21 March 2017, Version of Record 30 March 2017.

论文官网地址：https://doi.org/10.1016/j.eswa.2017.03.041