作者: M. Hariharan , Vikneswaran Vijean , R. Sindhu , P. Divakar , A. Saidatul
DOI: 10.1016/J.COMPELECENG.2014.01.010
关键词:
摘要: In recent years, various physiological signal based rehabilitation systems have been developed for the physically disabled in which electroencephalographic (EEG) is one among them. The efficiency of such a system depends upon processing and classification algorithms. order to develop an EEG or assistive system, it necessary effective algorithm. This paper proposes Stockwell transform (ST) analysis dynamics during different mental tasks. signals from Keirn Aunon database were used this study. Three classifiers employed as k-means nearest neighborhood (kNN), linear discriminant (LDA) support vector machine (SVM) test strength proposed features. Ten-fold cross validation method was demonstrate consistency results. Using method, average accuracy ranging between 84.72% 98.95% achieved multi-class problems (five tasks).