Differential evolution algorithm as a tool for optimal feature subset selection in motor imagery EEG

作者：

Highlights：

• We propose a novel hybrid method for feature selection using Differential Evolution for EEG signals.

• The proposed method achieved feature reduction by 90% and average classification accuracy of  95%.

• Evaluation of the proposed scheme was performed on BCI Competition III, dataset Iva.

• Results demonstrated promising performance compared with other evolutionary algorithms.

摘要

•We propose a novel hybrid method for feature selection using Differential Evolution for EEG signals.•The proposed method achieved feature reduction by 90% and average classification accuracy of  95%.•Evaluation of the proposed scheme was performed on BCI Competition III, dataset Iva.•Results demonstrated promising performance compared with other evolutionary algorithms.

论文关键词：Differential evolution,EEG,Feature selection,BCI,Motor imagery,CSP

论文评审过程：Received 7 December 2016, Revised 22 June 2017, Accepted 21 July 2017, Available online 22 July 2017, Version of Record 18 August 2017.

论文官网地址：https://doi.org/10.1016/j.eswa.2017.07.033