Abstract

Student satisfaction plays an important role in determining the quality, retention, and reputation of an institution. However, limited survey data can reduce the accuracy of predictive models. This study explores how Genetic Algorithm based data augmentation can improve dataset reliability and enhance analysis using LASSO and Ordinal Regression. By generating synthetic responses, GA expands the dataset while maintaining statistical accuracy, leading to better feature selection and ranking. LASSO Regression identified key factors influencing student satisfaction, such as career services, curriculum relevance, faculty support, and extracurricular activities, while Ordinal Regression pointed out that administrative inefficiencies negatively affect satisfaction levels. The results highlight that academic and career- related aspects have a greater impact on student satisfaction than infrastructure facilities. To enhance student experiences, institutions should focus on faculty mentorship, career counseling, and aligning the curriculum with industry needs. This study shows that GA-based data augmentation can significantly improve predictive modeling for student satisfaction analysis and offers practical recommendations for institutional development. Future research can incorporate machine learning techniques for more accurate predictions and tailored strategies to improve student success.

Authors

P. Priyadarshini, K.T. Veeramanju
Srinivas University, India

Keywords

Student Satisfaction, Genetic Algorithm, Ordinal Regression, LASSO Regression