Enhancement of Mahalanobis–Taguchi System via Rough Sets based Feature Selection

作者：

Highlights：

• Paper develops a new feature selection method for Mahalanobis–Taguchi System.

• New feature selection includes regularization or penalty for over-fitting.

• It also describes causal relationship between variables and classes.

• Results of proposed method are compared with those of other classification methods.

摘要

•Paper develops a new feature selection method for Mahalanobis–Taguchi System.•New feature selection includes regularization or penalty for over-fitting.•It also describes causal relationship between variables and classes.•Results of proposed method are compared with those of other classification methods.

论文关键词：Data mining,Mahalanobis Taguchi System,Feature Selection,Orthogonal Arrays,Rough Sets,Over-fitting,Regularization,Conditional probability,IF–THEN rules

论文评审过程：Available online 18 June 2014.

论文官网地址：https://doi.org/10.1016/j.eswa.2014.06.019