The Analysis of EEG Signal and Comparison of Classification Algorithms Using Machine Learning Methods

Abstract

In this paper we present comparison of classification models on EEG dataset to recognize the possibility of difference between right-handed and left-handed subjects. The research come out from the hypothesis, that it is possible to differentiate between different states of mind for left and right-handed people based on their EEG signal. In our research, there were used several machine learning methods like K-neighbors, support vector machines and decision tree classifier. These methods were explored to seek the difference between EEG signal of subjects using signal processing. The EEG data were obtained during reading various particular text samples with different meanings and from different areas. First chapters of this paper are devoted to the description of the experiment and data analysis with EEG signal processing. The final parts of the paper are evaluating our hypothesis from the previous parts of this paper with the use of various machine learning methods. We found correlations between left-handed people during reading the text samples. The correlations were significant while reading horror text sample.