Abstract: This paper introduces an emotion recognition system based on electroencephalogram (EEG) signals. EEG signals are classified into four emotional states happy, relax, sad and fear. We have used pre-processed dataset of EEG signals to build an emotion recognition system. To evaluate classification performance, Support Vector Machines is used and for feature extraction AR (Auto Regression) and FFT (Fast Fourier Transform) is used. This paper provides a methodology which is easy to understand and perceive emotion recognition process.

Keywords: EEG, FFT, AR, SVM