A Simple Framework for Predicting Student CGPA using Multi Regression Techniques

This paper proposes a framework for predicting a student's cumulative grade point average using the machine learning algorithm. A CGPA is a score that helps both academic institutions and employers to assess the capabilities of students. The proposed framework uses a dataset from Kaggle and applies regression techniques such as Linear Regression, Decision Tree Regression, Support Vector Regression (SVR), Gaussian Process Regression (GPR), and K-Nearest Neighbors (KNN) to predict CGPA based on the previous academic performance. The models were assessed using metrics such as accuracy, mean squared error (MSE), and R² scores. SVR and GPR outperformed the rest with accuracy of 98.58% and 98.56%, respectively, with less error rates. The research points out the possibility of using machine learning in identifying students who are at risk academically, which would allow timely interventions and improvement in educational outcomes. These findings provide a strong foundation for institutions to implement data-driven strategies that foster student success and meet higher standards of an institution. The model not only forecasts CGPA but also aids in detecting students who are academically at risk. This enables focused interventions, making it a valuable resource for evaluating students' academic abilities and identifying potential learning deficiencies.