Comparative Analysis of Naive Bayes and K-NN Approaches to Predict Timely Graduation using Academic History

Abstract

Graduation is an important benchmark in higher education accreditation and community assessment. This research seeks to ensure timely graduation for all students, considering the important role of higher education in this regard. To achieve this goal, data sets that cover academic performance during undergraduate and graduate studies are essential. The dataset includes details of universities, study programs, undergraduate and master’s GPAs, TOEFL scores, and duration of study. Utilizing classification techniques in data mining, specifically Naive Bayes and K-Nearest Neighbors (K-NN), a comparative analysis was conducted to accurately predict graduate students’ on-time completion. The graduation prediction process begins with data preprocessing, transformation, and division into training and testing sets. Next, the method is applied, and analysis is carried out to predict graduation outcomes. Experimental findings show that the Naive Bayes technique achieves an accuracy rate of 80%, while K-NN reaches 73%. Notably, Naive Bayes demonstrated superior efficacy in predicting on-time graduation. Efforts to improve accuracy require expanding data sets and diversifying variables. The findings of this research can guide academic institutions in implementing proactive measures to support students in completing their studies within the expected timeframe.