Mind Waves Time Series Analysis of Students’ Focusing and Relaxing Sessions

Abstract

Research uses electroencephalography (EEG) to study the reflection of emotional and physical activity on the mind’s behaviour. It allows the understanding of the mental status of students during educational sessions or patients during meditation sessions. Attention and Meditation are two EEG meters that reflect the mind focus and the mind calmness, respectively. The values of these meters across session time are considered as a time series data of a constant time unit. Current research considers that can discriminate the mind status according to the mind wave values at a specific time, whether an emotion like feeling happy is currently experienced or an activity like blinking is performed. This work proposes a new time series representation method to classify the gender of the student. This method considers the trend of time series values of mind wave meters during a specific activity. A composite feature vector of this representation is supplied to a machine learning model. The classification accuracy is boosted from 50% using traditional models to 70% using the proposed model. In addition, the results show that the accuracy of used classifier (decision tree) increases as the degree of polynomial fitting curve of the timer series vector increases, until a global beak then starts to decrease. The fitting degree is an important parameter in the proposed time series representation model .