Enhanced Student Retention In Open And Distance Education Through Effective Academic Performance Model Using Naïve Bayes And K-Nearest Neighbor Machine Learning Algorithms

Abstract

Improving student performance in an academic pursuit is one of the key concerns of institutions especially open and distance learning institutions where learners are separated from the institution by geographical region. The current observation of low-quality graduates from colleges and universities, particularly in open and distance learning, can be attributed to the lack of mechanisms that could help administrators at universities to forecast the academic achievement of the concerned students in the coming years. The goal of data mining in education is to create models, algorithms, and techniques for analyzing information gathered from learning environments to comprehend and enhance the learning process. The goal of this research is to identify patterns in the measures of academic achievement and how they relate to admission, high school, and personal information about the students. These findings can serve as a solid basis for customizing and enhancing the curriculum for open and distance learning to better suit the needs of individual students. Also, the research work identified factors that had a crucial influence on overall students’ performance. Hybridizing Naïve Bayes and K-Nearest Neighbor were used as Classifiers to develop a model for predicting the performance of students. The new model which is the hybridized model (combined Naïve Bayes and K-Nearest Neighbor) predicts better results than individual Naïve Bayes and K-Nearest Neighbor algorithms which shown itself as the best prediction and classification model.