A Comparative Study of Data Mining Techniques for Predicting Candidates’ Performance Based on High School Records

Abstract

Pre-admission entry requirements vary between higher education institutions. The pre-test performance of applicants in the entrance exam of a public engineering university is examined in this article based on their high school grades. Additionally, this study looked into the relationship between the scores obtained on high school exams and the entrance exam. Creating a prediction model to aid admissions committees in their work was the study's main objective. The candidates' pre-admission entry exam scores as well as their high school scores are included in the dataset.  To determine the association between these scores, a number of statistical analytic methods and machine learning algorithms, "Decision trees", "K-Nearest Neighbour, "Neural Networks," " and "Naive Bayes" were used. Accuracy metrics was applied to calculate the performance of the selected methods.. The 'Decision tree' and 'Neural network' models fared better than the other models, according to the results. The study concluded that the high school scores are significant predictors of pre-test performance and that machine learning models can be effective tools in predicting the performance of applicants in university entrance exams. The results of this study have applications for universities and admissions committees because they can be used to enhance selection procedures and spot potential high achievers.