Predicting learning behaviour of students: Strategies for making the course journey interesting

This paper focuses on improving higher education with the help of data mining techniques. The emphasis is on analysing specific attributes of students and predict their learning behaviour. A student can be a slow, medium or fast learner depending on his/her activeness in educational activities. Experimental results in this study show that decision tree classifier performs significantly well compared to nearest neighbour and naîve Bayesian algorithms. This paper also describes some strategies to seek students' attention to stay attached with classroom sessions and complete the course with full involvement.