Classification of Electroencephalogram (EEG) and graphology for students using Artificial Neural Network (ANN)

Abstract

Emotions define an individual's physiological conditions and are created subconsciously. The emotions can be happy or sad emotions. Emotions may have an impact on a person's handwriting. Because emotions have a link with cognitive functions, they have an influence on cognitive tasks such as writing. The emotion also can be identified through the features of handwriting. Handwriting is one of the unique characteristics to represent what is in our minds, to communicate with others. There are many features of handwriting such as baseline, slant, pressure, spacing, and others. So, the aim of this research is to make a scientific proven between the EEG pattern and graphology analysis based on emotion recognition. The EEG data and handwriting data are collected at the same time from the subjects. The Emotive is used to record the EEG data during the subjects are writing. Next, the signal will be processed using empirical mode decomposition (EMD) in order to remove any noise and unwanted data from the EEG signal where a clean EEG signal will be produced. Then, feature extraction for emotion identification is extracted through the power spectral density (PSD) of the subjects. Finally, this sample will be classified using an artificial neural network (ANN) where the accuracy for the happy emotion is 96.2% and for the sad emotion is 95.5%. Happy emotion will have an ascending baseline and higher value of Alpha power while sad emotion will have a descending baseline and the value of Alpha power is decreased compared to happy emotion. In conclusion, both EEG pattern and handwriting reflects the emotion of the subjects.