Novel associative classifier based on dynamic adaptive PSO: Application to determining candidates for thoracic surgery

作者：

Highlights：

• Novel associative classifier based on modified Particle Swarm Optimization (PSO).

• Uses local, global and personal learning; dynamic regions and adaptive parameters.

• Quality evaluation is done for individual rules as well as rule sets.

• Results show superior performance than fourteen state-of-the-art classifiers.

• Method is successfully applied to a practical medical domain problem.

摘要

•Novel associative classifier based on modified Particle Swarm Optimization (PSO).•Uses local, global and personal learning; dynamic regions and adaptive parameters.•Quality evaluation is done for individual rules as well as rule sets.•Results show superior performance than fourteen state-of-the-art classifiers.•Method is successfully applied to a practical medical domain problem.

论文关键词：Associative classification,PSO,Rule quality

论文评审过程：Available online 16 July 2014.

论文官网地址：https://doi.org/10.1016/j.eswa.2014.06.046