Predictive Modeling of Student Academic Performance in External Examinations

In a country like India, where students constitute a major part of the population, there is a need to predict a student’s performance. The need for analysing students’ academic performance plays an important role in the education sector because the insights gained from this prediction can be used to redesign the curriculum, which helps in meeting the needs of individual students. Educational data mining has been an emerging field to improve educational institutions. Hence, the usage of machine learning techniques in educational data mining has increased rapidly. Accurately predicting students’ academic performance is essential for various stakeholders, like instructors, students, and institutions. This study aims to predict a student’s performance in external exams using various machine learning techniques. Various metrics, like internal exam scores, previous semester scores, and the attendance of a student, are being considered for predicting GPA in the current semester. Because of the unbalanced nature of the data, it is challenging to develop a high-accuracy model. For this prediction, a variety of machine learning methods can be applied, including Random Forest (RF), K-Nearest Neighbour (kNN), Naïve Bayes (NB), Gradient Boost (GB), Decision Tree (DT), Support Vector Machine (SVM), and Logistic Regression (LR). Ensemble methods can demonstrate high accuracy in model training. By analysing the results produced by our proposed model, valuable insights can be drawn that benefit students’, teachers, and educational administrators.