Classification of video lecture learners' cognitive and negative emotional states using a Bayesian belief network


Xiaomei Tao, Qinzhou Niu, Mike Jackson, Martyn Ratcliffe




In general, Intelligent Tutoring Systems (ITS) fail to take account of the emotional and cognitive states of the students who use them. This paper explores the relationship between emotion and cognition when students learn via the medium of video lectures. A cognitive emotional model was constructed to determine the student's cognitive and emotional state while watching an instructional video. This model was a Bayesian belief network (BBN) model. With the method of ten times 10-fold cross-validation, evaluation results showed that the Bayesian network classifies the emotion state with 60% accuracy and classifies both the emotion and cognitive state with 48.82% accuracy. This model provides an emotional and cognitive states recognition solution for video lecture learners in a non-intrusive way with low cost.