Classification of human emotions from EEG signals using SVM and LDA Classifiers

Abstract

Emotion Detection has been a topic of great research in the last few decades. It plays a very important role in establishing human computer interface. We as humans are able to understand the emotions of other person but it is literally impossible for the computer to do so. The present work is to achieve the same as accurately as possible. Emotion detection can be done either through text, speech, facial expression or gesture. In the present work the emotions are detected using Electroencephalography (EEG) signals. EEG records the electrical activity within the neurons of the brain. The main advantage of using EEG signals is that it detects real emotions arising straight from our mind and ignores external features like facial expressions or gesture. Hence EEG can act as real indicator of the emotion depicted by the subject. We have employed Independent Component Analysis (ICA) and Machine Learning techniques such as Support Vector Machine (SVM) and Linear Discriminant Analysis (LDA) to classify EEG signals into seven different emotions. The accuracy achieved with both the algorithms is computed and compared. We are able to recognize seven emotions using the two algorithms, SVM and LDA with an average overall accuracy of 74.13% and 66.50% respectively. This accuracy was achieved after performing a 4-fold cross-validation. Future applications of emotion detection includes neuro-marketing, market survey, EEG based music therapy and music player.