Prediction of Student’s Academic Performance using k-Means Clustering and Multiple Linear Regressions

Abstract

In today's educational system, performances of students are mainly based on tests, assignments, attendance, quizzes and final examination. It is at the end of this exercise that a minimum mark is determined on which promotion will be based. There is need to identify factors that lead to a student's success or failure. This will allow the teacher to provide appropriate coWIBelling and focus more on such factors. Hence, a model for forecasting student's performance academically is of a pronormced significance, therefore, data mining techniques in classifying and forecasting the academic performance of students was put into application in this research study. k-means clustering and Multiple Linear Regression (1.1LR) were used for assessing student's performance. The results showed that student's test seores, quiz and assignment were the major factors that could be used in predicting academic performance of students. Also, two clusters were derived with the use of elbow method to group all the students into clusters.