Understanding Internet Usage Patterns and Behaviors Among College Students: A Machine Learning Approach

This study explores internet usage patterns and behaviors among students enrolled in Information Technology Education (ITE) programs in the province of Bohol, Philippines. Despite the widespread availability of Internet resources, there is limited understanding of how ITE students utilize the Internet for academic and personal purposes. This research aims to identify key patterns and factors influencing internet usage frequency at school, home, and other locations. Utilizing machine learning techniques and exploratory data analysis (EDA), the study analyzed data from a structured questionnaire administered to ITE students through Google Forms. Classification models, including Random Forest Classifier, Decision Trees, and Support Vector Classifier, were employed to predict internet usage levels based on demographic and behavioral data. Results indicate that ITE students exhibit high levels of internet usage, with significant correlations between the frequency of internet usage at school, home, and for various purposes with overall internet usage levels. Among the models, SVC achieved the best performance, with an accuracy of 99.53%. In comparison, Random Forest Classifier attained 92.63% accuracy, while Decision Trees performed at 84.02% accuracy. These insights can inform educational institutions and policymakers in developing strategies to enhance internet accessibility and digital literacy among students, ultimately improving educational outcomes and digital inclusion.