Classification of two class motor imagery task using Jaya based k-means clustering

The brain-computer interface (BCI), identify brain patterns to translate thoughts into action. The identification relies on the performance of the classifier. In this paper identification of electroencephalogram (EEG) based BCI for motor imagery (MI) task is done through asynchronous approach. Transferring the brain computer interface (BCI) from laboratory state to real time application desires BCI to be functional asynchronously exclusive of any time constraint. In this case, ongoing brains activities are analyzed continuously irrespective of state i.e. control or non-control. The k-means is the most commonly used and one of the best clustering technique. With k-means, there is a possibility of convergence to the local minima. In literature, it is reported that k-mean modified with optimization techniques have able to address this problem by avoiding premature convergence of local minima. In this article, we have used algorithm-specific parameter-less Jaya optimization to select the initial cluster for k-means. Though enhancement in the accuracy is marginal, it is the speed which makes Jaya based k-means to be an obvious choice for the problem under consideration, i.e. classifying motor imagery task. Supremacy of the proposed Jaya based k-means among the considered algorithm is validated through Friedman test.