Utilizing Cluster Analysis and Discriminant Analysis for Data Classification and Academic Performance Prediction

Abstract

This study aims to utilize cluster analysis and discriminant function analysis to classify student data into two groups: high and low academic performance, and to evaluate the accuracy of the cluster classification. The study adopted a descriptive research design, and the sample consisted of a random sample of 62 students. Both cluster analysis and discriminant analysis were applied to the data. The results of the cluster analysis indicated that Aptitude Test scores and Achievement Test scores played a significant role in classifying cases into clusters. In addition, the results of the discriminant function analysis showed that Wilks’ Lambda was statistically significant, indicating the discriminant function's ability to distinguish between the two groups. Furthermore, the assessment of classification accuracy revealed that the classification rate based on cluster analysis was very high, confirming the accuracy and effectiveness of the classification process. The study recommends applying cluster analysis to classify cases into homogeneous or closely related clusters within each group.