Bullying Prevention in Sarawak: An Ethical Predictive Model Leveraging Community and Institutional Intelligence

Abstract

This conceptual paper proposes an integrated framework for deploying predictive analytics within early intervention systems for school bullying. Moving beyond reactive approaches, the framework describes a data-driven platform designed to identify latent risk factors and forecast the probability of bullying incidents within educational institutions. The developing Sistem Intervensi Awal & Pemantauan Buli Sarawak (SIAP-Buli) is used as a case study to illustrate practical design, multimodal data sources, and ethical governance. Grounded in a systematic review of predictive modelling in social science and public-health informatics, the framework synthesises evidence from Scopus-indexed literature and conference proceedings to build a theoretical model. Key analytical components include Geographic Information Systems (GIS) for spatiotemporal mapping, supervised and unsupervised machine learning for pattern discovery, and interactive dashboards for stakeholder decision-making. The conceptual model identifies that fusing structured quantitative records (incident logs, school metrics) with unstructured qualitative inputs (free-text surveys, anonymous tips, sentiment indicators) is essential for robust prediction. It posits that targeted, data-informed interventions—guided by probabilistic risk stratification—can materially reduce bullying prevalence. The paper offers original theoretical groundwork for operationalising predictive analytics in school safeguarding, with particular emphasis on socio-cultural and geographic complexities found in East Malaysia. It outlines governance measures to mitigate bias and protect student privacy, providing a scalable blueprint for other jurisdictions.