Model for Analyzing Psychological Parameters Recommending Student Learning Behaviour using Machine

Abstract

One of the main objectives of Educational Data Mining (EDM) is to improve the education system to increase student retention, help students to score high and attain holistic development. The purpose of the study is to analyze the psychological parameters of students to predict their intellectual performance and generate recommendations to be utilized by institutes and students to improve academic performance. This study performs matrix factorization using single value decomposition (SVD) to predict missing parameters related to the psychological behaviour of students with root mean square equals 0.059 and uses the userbased collaborative filtering technique to predict their grade with RMSEA as 0.055. It makes use of decision tree (ID3) algorithm for generating decision rules that produces results with an accuracy of 76% and provides suggestions on how to improve learning by changing the psychological behaviour of students. The results showed that three parameters of personality (namely conscientiousness, openness and need for cognition), six of motivation construct (intrinsic motivation, optimistic, goal orientation, concentration, locus of control and self-efficacy), five of self-regulatory learning strategies construct (rehearsal, elaboration, metacognition, peer learning, time/study management) highly impacted academic performance in positive way. Students belonging to upper and middle socioeconomic status avail more from learning facilities. Also, learning the in-depth knowledge of the topic enhance student intellectual performance. It is noted that social integration and academic integration help students to learn the subject matter in friendly environment and reduces depression. The key findings highlight the parameters positively impacting students’ intellectual performance. This help in improving students’ intellectual performance which further addresses student retention, progress and employability.Keywords: Academic Performance, Single Value Decomposition, Student, Educational Data Mining, Prediction, Recommendation, Collaborative Filtering, Decision Tree.