Predicting Digital Addiction Patterns with Machine Learning for Personalised Mental Health Support - Journal of Social Computing Volume 6 Number 4

Aug 4, 2026

The exponential growth of social media has intensified concerns about digital addiction and its mental health impacts, especially among younger populations. Existing digital health tools can offer support but often overlook the predictive value of structured behavioural data. This study proposes a machine learning framework using 3200 anonymised self-reports covering screen time, social media engagement, sleep duration, and mental health indicators. CatBoost outperforms alternatives, while linear regression and structural equation modelling clarify interpretable and mediating pathways through anxiety and depression.

LanguageEnglish

CategoryHealthcare

FormatPDF

Pages19

Reading Time48 min

AccessFree to download

Introduction

Background: mental health and digital addiction

Vulnerability of Generation Z and social media mechanisms

Research gaps in prediction and mitigation

JOURNAL OF SOCIAL COMPUTING Volume 6, Number 4, December 2025 ISSN 2097-5015 07/07 pp359−377 DOI: 10.23919/JSC.2025.0020 Predicting Digital Addiction Patterns with Machine Learning for Personalised Mental Health Support Ayodeji Olusegun Ibitoye*, Shaiyini Ravindran, Oluwaseyi Funmi Afe, and Adeyinka O. Abiodun Abstract: The exponential growth of social media has heightened concerns about digital addiction and its mental health consequences, particularly among younger populations. Existing digital health tools, including conversational agents and large language models, offer real-time support but often neglect the predictive value of structured behavioural data. This study introduces a machine learning framework to assess digital addiction risk using 3200 anonymised self-reports comprising screen time, social media engagement, sleep duration, and mental health indicators. Across multiple models, categorical boosting (CatBoost) achieves the highest performance (precision = 85.4%, receiver operating characteristic-area under the curve (ROC-AUC) = 0.93), outperforming extreme gradient boosting (XGBoost) and graph neural networks (GNN) . A linear regression model provides interpretable correlations between behavioural variables and addiction risk. Structural equation modelling (SEM) reveals that anxiety and depression mediate the relationship between digital behaviours and addiction risk, offering causal insights into these pathways. Feature importance analysis identified excessive screen time, frequent social media checking, and reduced sleep as the most influential predictors. To translate findings into practice, K-means clustering generated behavioural risk profiles, enabling personalised, data-driven recommendations. While clinical validation remains a next step, this framework demonstrates how predictive modelling and clustering can inform scalable, noninvasive digital health interventions. By integrating machine learning with causal modelling and personalised intervention design, this study advances computational approaches to digital addiction and contributes to the broader discourse on artificial intelligence applications in mental health and social computing. Key words: digital addiction; machine learning (ML); mental health; personalised recommendations; social media behaviour; predictive modelling; digital wellness 1 Introduction Mental health is a cornerstone of overall well-being, significantly influencing how individuals think, feel, and function in their daily lives. It encompasses conditions ranging from anxiety and depression to complex Ayodeji Olusegun Ibitoye is with School of Computing and Mathematical Sciences, University of Greenwich, London, SE10 9LS, UK. Email: [a.o.ibitoye@greenwich.ac.uk](a.o.ibitoye@greenwich.ac.uk). Shaiyini Ravindran is with School of Human Science, University of Greenwich, London, SE10 9LS, UK. E-mail: Shaiyini.Ravindran@greenwich. [ac.uk](ac.uk). Oluwaseyi Funmi Afe is with Department of Computer Science, Lead City University, Ibadan 200255, Nigeria. E-mail: [afe.seyi@lcu.edu.ng](afe.seyi@lcu.edu.ng). Adeyinka O. Abiodun is with Department of Computer Science, National Open University of Nigeria, Abuja 900241, Nigeria. E-mail: [aabiodun@noun.edu.ng](aabiodun@noun.edu.ng). * To whom correspondence should be addressed. Manuscript received: 2025-04-05; revised: 2025-08-31; accepted: 2025-09-08 © The author(s) 2025. The articles published in this open access journal are distributed under the terms of the Creative Commons Attribution 4.0 International License ([http://creativecommons.org/licenses/by/4.0/](http://creativecommons.org/licenses/by/4.0/)). 360 Journal of Social Computing, December 2025, 6(4): 359−377 disorders like schizophrenia and bipolar disorder[1]. The World Health Organisation (WHO) defines mental health as the capacity to cope with life’s stresses, work productively, and contribute to society[2]. Despite its critical role, mental h

FAQs

What data and variables are used to predict digital addiction risk?

The framework uses 3200 anonymised self-reports with screen time, social media engagement, sleep duration, and mental health indicators.

Which machine learning model achieves the best predictive performance?

CatBoost achieves the highest performance, with precision of 85.4% and ROC-AUC of 0.93 in the reported results.

How do anxiety and depression influence digital addiction in the study?

Structural equation modelling indicates that anxiety and depression mediate the relationship between digital behaviours and addiction risk, providing causal pathway insights.

Aug 4, 2026

Upload by Theodora