Design and Analysis of Students Academic Performance Prediction System Using Improved Machine Learning Methodologies

Abstract

Academic achievement, social justice, and economic progress all depend on having access to higher education.  But dropout rates are a big problem for schools all throughout the world.  A number of factors, including socioeconomic status, contribute to the large variation in dropout rates among nations.  To improve retention rates and implement effective interventions, at-risk students must be identified early.  This study uses a range of machine learning techniques to predict whether students will succeed academically or drop out.  We assessed the demographic, socioeconomic, academic, social, and macroeconomic characteristics of  students enrolled in  distinct majors   The dataset includes 35 attributes, including special educational needs, gender, scholarship status, age at enrolment, debt status, tuition fee status, marital status, application mode, course, attendance type, prior qualifications, nationality, parental qualifications and occupations, and curricular unit performance.  The data was pre-processed by identifying relevant classes and attributes, eliminating outliers using the Interquartile Range (IQR) method, and removing negative correlations from features.  After normalizing the dataset using Standard Scaler, we divided it into two sets: a training set, which accounted for 67% of the total, and a testing set, which included the remaining 33%.  Grid search was used to optimize the hyperparameters.  Six classification algorithms—SVM, Decision Tree, Random Forest, Naive Bayes, K-Nearest Neighbors (KNN), and Logistic Regression—were used to create prediction models.  The SVM model was shown to have the best accuracy, precision, recall, and F1-score.  Compared to Naive Bayes, KNN, and Decision Trees, Random Forest and Logistic Regression performed better.  The results demonstrate the efficacy of Random Forest, SVM, and Logistic Regression models in forecasting students' school departure times.  This study highlights the importance of machine learning in improving educational administration and raising student achievement by giving schools useful tools for early risk assessment and tailored intervention tactics.