International Journal of Scientific Research and Engineering Development

Background: Previous studies on screen time and its health impacts have primarily focused on specific age groups, such as children or adults, and observed only a single health outcome, such as obesity or sleep disorders, by finding a direct correlation between increased screen exposure and higher BMI. The primary goal of this project is to analyse the relationship between screen time and its physical, mental, emotional, social, and academic side effects across all age groups. By creating a unique dataset that links screen time with health indicators, the project aims to develop a machine learning model to predict and mitigate the risks associated with excessive screen time. Methods: The dataset consists of 999 records (rows) and 16 columns, with 8 input features and 8 output features. The input features capture digital usage behaviours such as social media, gaming, and streaming, while the output features represent health impacts like anxiety, sleep disruption, and physical pain. This size allows for a strong analysis of the relationship between screen time and health outcomes across different age groups. Result: By clustering algorithms such as Self-Organizing Maps (SOM), KMeans, and Fuzzy C Means, we were able to categorize individuals based on their screen time patterns and associated health risks. The Heatmaps were applied to visualize the clustering results of Self-Organizing Maps (SOM). These Clustering diagrams effectively display the distances between nodes, indicating areas of higher similarity. The darker regions represent clusters with minimal distances, signifying greater user similarity. Conclusion: This project used advanced machine learning techniques (SOM, K-Means, Fuzzy C-Means) to analyse screen time's impact on health, revealing strong links between excessive digital device usage and issues. The SOM algorithm achieved 96.23% accuracy in predicting health risks, while clustering methods helped categorize users by risk levels, enabling proactive screen time management.

Ch.Rajasekhara Rao, K.Eswara Rao, K.Sravanthi, Boddepalli Alekya,"Predicting Public Health Consequences Using Digital Phenotyping Based on Screen Time: A Machine Learning Perspective" International Journal of Scientific Research and Engineering Development, V9(3): Page(1754-1763) May-June 2026. ISSN: 2581-7175. www.ijsred.com. Published by Scientific and Academic Research Publishing.