Using Decision Tree to Figure out Relationship between Physiological Signals and Personality Traits for Students Learning Programming

Abstract

This research aims to solve the problem of students’ learning difficulties in learning programming courses without a suitable learning strategy. This usually results in poor learning results. Existing research showed that there is a relationship between personality traits and suitable learning methods. According to Duff, learners with strong extraversion and openness tend to favor the deep learning approach. Those with strong neuroticism and agreeableness tend to favor the surface learning approach, while those with strong extraversion and conscientiousness tend to have a lower degree of neuroticism. In this research, the goal is to use physiological signals to measure personality traits. For the goal, random forest training models are adopted. The research results are summarized as follows. High skewness of voice frequencies indicates high Openness to Experience and vice versa. Low standard deviation of voice frequencies indicates high Extraversion and vice versa. Low skewness of voice frequencies and low standard deviation of GSR indicates low Emotion Stability. Low skewness of voice frequencies and high standard deviation of GSR indicates high Emotion Stability. High skewness of voice frequencies and low standard deviation of GSR indicates low Emotion Stability. High skewness of voice frequencies and high standard deviation of GSR indicates low Emotion Stability. Low skewness of voice frequencies and low variance of GSR indicates low Conscientiousness. Low skewness of voice frequencies and high variance of GSR indicates high Conscientiousness. High skewness of voice frequencies indicates high Conscientiousness. Low skewness of GSR and low skewness of voice frequencies indicate low Agreeableness. Low skewness of GSR and high skewness of voice frequencies and low standard deviation of voice frequencies indicate high Agreeableness. Low skewness of GSR and high skewness of voice frequencies and high standard deviation of voice frequencies indicate low Agreeableness. High skewness of GSR and low skewness of voice frequencies indicates high Agreeableness; high skewness of GSR and high skewness of voice frequencies indicates low Agreeableness. Based on the measured personality traits, it is then possible to choose the appropriate learning strategies for better learning performance.