Prediction of Student’s Academic Performance Using Linear Regression

Abstract

Recently, there has been a tremendous increase of failure rate in higher institutions. Often times, students of tertiary institutions drop-out of school or end up with certain classes of degrees that are far lesser than their ideal intellectual capabilities. As fresh students enroll into higher institutions, some may not even know what a class of degree is, and the level of work that is likely to place them in a particular class of degree. Predicting their academic performance in this wise becomes necessary to ascertain the future of their performances in order to make informed decisions. Traditionally, we find out “who failed in an examination” or who “passed in an examination”. However, predictions usually finds out who is likely to fail and who is likely to pass an examination and to what extent the failure or pass could be. Hence, predictions serve as counseling tool for the students to either improve their work rate or maintain their rate to achieve higher performance in subsequent examinations. In this paper, we predicted the academic performance of Benue State University students using linear regression. Linear regression assumes a linear relationship between the dependent variables (x) and an independent variable (y) such that the slope (predicted) can be specifically calculated from a computational linear combination of the variables x and y. Our proposed system correctly predicted the performance of mathematics/computer science students of the Benue State University with an accuracy of up to 100%.