K-means Clustering Analysis for EEG Features of Situational Interest Detection in Classroom Learning

This paper proposes a method to detect situational interest in classroom learning using k-means algorithms. The developed algorithm in this paper had been tested on features from ten students who experienced mathematics learning in a classroom. The subjects were given 21 min of Laplace lecture presentation with some interesting elements introduced. Electroencephalogram (EEG) signal was preprocessed and decomposed using Fast Fourier Transform. The mean power for each sub-frequency band was served as input to the k-means algorithm. Results showed that EEG features can be successfully clustered in the alpha frequency band at the frontal region when visual-auditory stimuli are introduced to the subjects. © 2021, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.