BRAIN COMPUTER INTERFACE IMPLEMENTATION USING CLUSTERING TECHNIQUES

Brain Computer Interface (BCI) generally utilizes non-invasive EEG signals in order to detect intended movement. Normally all EEG electrode receive brain activity on scalp surface which is superimposition of different brain activity. As the number of channel is high we need to reduce it in order to properly detect movement related activity. In this work we have utilized different clustering algorithms to reduce insignificant channels. We have extracted features using statistical methods Principal Component Analysis (PCA). And the classification is done using unsupervised learning algorithm in order to improve generalization capability. The proposed approach reduces false positive rate and it also shows different person have different activity region.