Analysis on EEG signal with machine learning

In this paper, research on electroencephalogram (EEG) is carried out through principal component analysis (PCA) and support vector machine (SVM). PCA is used to collect EEG data characteristics to discriminate the behaviors by SVM methodology. The actual EEG signals are obtained from 18 experimenters who raised hands with meditation and actual movement during the experiments. The 16-channel data from the experiments form one data set. In order to get principal component of EEG signal, 16 features are considered from each channel and normalized. Simulation results demonstrate that two behaviors-i.e., raising hands and meditation-can be clearly classified using SVM, which is also visualized by a 2-dimensional principal component plot. Our research shows that specific human actions and thinking can be efficiently classified based on EEG signals using machine learning techniques like PCA and SVM. The result can apply to make action only with thinking.