ENHANCING E-LEARNING THROUGH STUDENT PERFORMANCE PREDICTION USING ML ALGORITHMS

Abstract

In the current era of data-driven decision-making, educational institutions are employing cutting-edge methodologies to assess and improve the academic performance of students. Predictive analytics, which employs machine learning methodologies, is an effective instrument for identifying patterns in educational data and forecasting future outcomes. The objective of this research is to develop a method for predicting the grades of students through the use of supervised machine learning algorithms. The dataset that was employed to develop the system encompasses academic and socioeconomic factors that influence students' performance. Data preprocessing entails the encoding of categorical variables, the removal of missing values, and the selection of critical features to improve the performance of the model. The optimal model is determined through the comparison and contrast of a variety of machine learning algorithms, including XGBoost, Linear Regression, Support Vector Machine (SVM), and K-Nearest Neighbors (KNN). The performance of these models is evaluated using a variety of metrics and visual aids, such as MAE, MSE, R2 Score, accuracy, and confusion matrices. The model that achieved the highest R2 score was the winner. This system can be employed by educational institutions to enhance learning outcomes by predicting student performance, identifying at-risk students at an early stage, and taking the necessary action.