A misclassification cost risk bound based on hybrid particle swarm optimization heuristic

作者：

Highlights：

• Proposed a misclassification cost risk bound (MCRB) for binary linear classification problems.

• Developed a hybrid particle swarm optimization procedure to solve MCRB.

• Using simulated and real-world datasets, test the MCRB bound for several linear and non-linear cost sensitive classifiers.

摘要

•Proposed a misclassification cost risk bound (MCRB) for binary linear classification problems.•Developed a hybrid particle swarm optimization procedure to solve MCRB.•Using simulated and real-world datasets, test the MCRB bound for several linear and non-linear cost sensitive classifiers.

论文关键词：Cost sensitive classification,Misclassification cost bound,Genetic algorithms,Particle swarm optimization

论文评审过程：Available online 29 August 2013.

论文官网地址：https://doi.org/10.1016/j.eswa.2013.08.045