A novel multi-objective medical feature selection compass method for binary classification

作者：

Highlights：

• Medical feature selection requires to take several objectives into account to comply with practitioners' needs.

• Medical classifiers must be accurate and guarantee the quality of predictions while considering the less features as possible

• The compass method allows to explore different search spaces that are relevant according to practitioners' needs.

• The best relevant medical subset of features eventually is the one which maximises accuracy and AUC scores.

摘要

•Medical feature selection requires to take several objectives into account to comply with practitioners' needs.•Medical classifiers must be accurate and guarantee the quality of predictions while considering the less features as possible•The compass method allows to explore different search spaces that are relevant according to practitioners' needs.•The best relevant medical subset of features eventually is the one which maximises accuracy and AUC scores.

论文关键词：Genetic algorithm application,Multi-objective feature selection,Extreme learning machine,Machine learning,Cardiology,Medicine

论文评审过程：Received 6 September 2021, Revised 15 January 2022, Accepted 6 March 2022, Available online 9 March 2022, Version of Record 12 March 2022.

论文官网地址：https://doi.org/10.1016/j.artmed.2022.102277