Predicting Academic Performance of Students Using Machine Learning Models

The NEP- 2020 of India has reemphasized the need of delivering quality education in order to decrease the drop out rate of students by the HEI. In post-cover era, retention of the students specially in Engineering and Management disciplines was more prominent. This paper is an attempt to empower HEI predict students’ performance/ potential drop out using ML algorithms based on six factors namely; Family size, Study time, Time-spent on extra-curricular activities, Absenteeism, Time spent on Internet, and Health.Three ML algorithms namely, KNN, Random Forest and Decision Tree have been implemented and compared on efficiency and accuracy criteria so as to attain at the top algorithm for subject prediction use case. The paper’s findings on academic performance prediction are likely to help educational institutions and the learners by the way of early identification of potential drop out and implementation of corrective measures. A small data set of 390 students on various predictors has been utilized for ML algorithms implementation.This paper seeks to address the problem of quality education delivery system in India and elsewhere. It is pilot research on academic performance monitoring with an aim to reduce dropout rate of students. This research work shall pave way for further scaled up research work implementing various algorithms on larger datasets to make a use case for education industry wide implementation potential.