A Hybrid Machine Learning Framework For Analyzing The Impact Of Social Media On Students’ Academic Performance

Abstract

In today’s educational landscape, social media plays a dual role—serving as a tool for collaboration and communication while also posing risks to student well-being and academic focus. This study introduces a hybrid machine learning framework designed to assess both academic success and behavioral risks associated with social media usage. Using pre-admission academic records and behavioral indicators such as sleep quality, mood disorders, and stress levels, the framework operates in two stages: first predicting academic success and then evaluating whether successful students are vulnerable to negative social media impact. Multiple models—including Random Forest, Support Vector Machine, Neural Network, Gradient Boosting, and an Ensemble approach—were tested, with Gradient Boosting achieving the highest accuracy (99%) across both stages. The results demonstrate the framework’s potential for early intervention, enabling institutions to proactively identify at-risk students and implement targeted support strategies. This approach offers a scalable, data-driven solution for enhancing student outcomes and promoting digital wellness in academic environments.