Predicting Online Vulnerability among Children and Youth: A Machine Learning Approach with Decision Trees and Naive Bayes Algorithms

Social media usage has surged among children and adolescents, leading to concerns about their online behavior and exposure to potentially embarrassing or harmful situations. This study aims to utilize Machine Learning (ML) techniques to predict the susceptibility of children and adolescents to such situations based on their social media usage patterns. A dataset was created through a survey conducted with users aged 9 to 14 in the city of Coimbatore, Tamil Nadu, which captured various aspects of their online behavior. The ML models used in this research were Decision Tree (DT) and Bayesian Network (BN), applied to analyze the data and identify behavior patterns. The key contributions of this research are twofold: first, it provides a detailed analysis of the social media behavior of children and adolescents in Coimbatore; second, it demonstrates the effectiveness of ML techniques in predicting online risks for this vulnerable group. By identifying behavior patterns, this study offers important insights into the online experiences of young people in Tamil Nadu, contributing to the development of more effective online safety measures and policies tailored to local needs.