Emotional State Classification in Machine Learning for EEG Signals

Abstract

Emotional health is very important to build our life. Emotional health is equally as important as our physical health. Dealing with our emotions is a difficult task because we can’t see them. To analysis emotional state is also the interested field of the researcher. We can test the brain wave to analysis the emotion using electro-encephalography (EEG) signals. There are many kinds of emotions. We propose to classify the human brain wave for happy, disgust, surprise, anger, sad and fear. To classify six types of human emotion needs data annotation, feature extraction, feature selection and classification methods by analyzing electroencephalography (EEG) signals. We propose time frequency domain analysis to EEG signals and the result is filtered by finite impulse response (FIR). The pure signal is label with logistic regression machine learning algorithm to classify the feature of the signals. We propose the good method for FIR and combine with logistic regression to show the high accuracy of the signal class