Impact of Mobile Device Usage on User Behaviour: Predictive Analytics and Digital Health Perspectives

Abstract

Mobile devices have transformed communication, information access, and consumer behavior, but excessive usage raises concerns about smartphone addiction. Despite extensive research, challenges persist in accurately predicting and mitigating the negative effects of excessive smartphone use. Traditional methods often struggle with biases, small sample sizes, and the inability to capture real-time behavioral patterns. This paper analyzes mobile usage patterns and employs machine learning models, including Random Forest Classifier and Multilayer Perceptron, to predict addiction levels with high accuracy. Using ANOVA for feature selection, the research enhances prediction reliability. A web application, developed with Streamlit and Python, provides real-time feedback and recommendations for healthier usage. These insights help businesses, policymakers, and health professionals promote digital well-being in an increasingly mobile-centric world.