Model Development in Predicting Academic Performance of Students Based on Self-Efficacy Using K-Means Clustering

Abstract

Many prediction models have been developed using data mining tools in order to predict the performance of the students at the early stage. The academic performance of higher education students commonly was predicted based on their results in the end of the previous semester or during the semester like test score or mid-term exam. However, there is lack of models that emphasize the use of data related to student’s behaviour for predicting the academic performance. Therefore, the aim of this study is to investigate the use of self-efficacy behaviour data to predict the academic performance of students using principal component analysis (PCA) and k-means clustering (KMC). This study focuses on the first part of the prediction which is model development. In the model development phase, the number of variables were reduced from 20 into two by using PCA. The scores for the first two principal components were clustered by using KMC. The results show that the scores can be clustered into three main clusters to differentiate the student’s self-efficacy behaviour. Next research will investigate the underlying causes of the clusters in order to predict the risky students.