Prediction of college student academic performance using data mining techniques.

• Main Authors: Abd Jalil, Azura, Mustapha, Aida, Santa, Dzulizah, Zain, Nurul Zaiha, Radwan, Rizalina
• Format: Conference or Workshop Item
• Language: English; English
• Published: 2013
• Online Access: http://psasir.upm.edu.my/id/eprint/31371/1/ID%2031371.pdf; http://psasir.upm.edu.my/id/eprint/31371/

For every learning institution whether schools, colleges or universities, high student success rate is the strategic goal that mirrors the reputation of an institution. However, case-by-case analysis is a daunting task and does not take into account data from a number of previous semesters or trend of overall performance in a particular semester. This study attempts to predict the success rate of students’ academic performance by analyzing their examination results to secure a place at college level for the subsequent semester. The classification algorithms used are the Decision Tree, Naïve Bayesian, and Multilayer Perception with the highest classification accuracy by the Naive Bayes algorithm with accuracy of 95.3%. High accuracy demonstrates the ability of data mining classification task in helping the institutions to predict student performance and to identify group of weak students. This will in turn help to improve their performance at much earlier stage.