Effect of the Multiple Intelligences in multiclass predictive model of computer programming course achievement
Abstract
This paper proposes the measurement of Multiple Intelligences (or MI) value as a co-determiner combine with the traditional academic achievements in predicting student performance in taking a computer programming course. The effectiveness of MI in such a predictive model is tested on three machine learning algorithms: Artificial Neural Network, Support Vector Machine, and the classic Naïve Bayes. Using three different validation schemes: 2, 5, and 10-folded cross validations, the results show that the Mi-inclusive model significantly helps to improve the accuracy of predicting students' performance. These are divided into three class: good, average, and poor achievement.
Authors: Unhawa Ninrutsirikun, Bunthit Watanapa, Chonlameth Arpnikanondt, Naphongthawat Phothikit
Published in: IEEE Region 10 Conference (TENCON) (2016)