An approach to EEG-based gender recognition using entropy measurement methods
作者：
Highlights：
• It was possible to use EEG signals for gender recognition. The highest recognition rate in this study was up to 0.998 based on a combination of FE feature and vote method, which could meet the needs of daily applications.
• The effect of the bagging ensemble method depended on the selection of the base classifiers and number of base classifiers.
• The GBDT method could obviously improve the performance of the detector.
• The vote classifier improved performance significantly.
摘要
•It was possible to use EEG signals for gender recognition. The highest recognition rate in this study was up to 0.998 based on a combination of FE feature and vote method, which could meet the needs of daily applications.•The effect of the bagging ensemble method depended on the selection of the base classifiers and number of base classifiers.•The GBDT method could obviously improve the performance of the detector.•The vote classifier improved performance significantly.
论文关键词：Gender recognition,Electroencephalogram (EEG),Entropy,Ensemble classifier,Receiver operating characteristic (ROC),Area under ROC curve (AUC)
论文评审过程：Received 14 June 2017, Revised 27 August 2017, Accepted 30 October 2017, Available online 4 November 2017, Version of Record 6 December 2017.
论文官网地址：https://doi.org/10.1016/j.knosys.2017.10.032
