Spatial Analytics And Machine Learning Integration For African Swine Fever Management: A Predictive And Prescriptive Decision Support Framework

Abstract

Background: African Swine Fever (ASF) has severely disrupted the Philippine hog industry since 2019, resulting in the culling of over 300,000 pigs and threatening national food security. Current management strategies are often hindered by inadequate monitoring and weak biosecurity protocols. Objectives: This study aims to develop and introduce a web-based Decision Support System (DSS) that integrates Geographic Information Systems (GIS), Density-Based Spatial Clustering of Applications with Noise (DBSCAN), and Random Forest Regressor algorithms to effectively identify disease hotspots and forecast future outbreaks. Method: A mixed-methods developmental design was utilized to construct the system. The DSS was subsequently evaluated based on ISO/IEC 25010 standards by a diverse panel of 258 stakeholders comprising policymakers, IT professionals, and veterinarians alongside comprehensive technical audits. Results: The system demonstrated high efficacy in distinguishing dense infection clusters from isolated "noise" cases using DBSCAN, while the Random Forest model generated accurate monthly forecasts for ASF cases and mortality. Stakeholders rated the system's functional dimensions from "Acceptable" to "Highly Acceptable," though technical audits identified critical areas requiring performance optimization and cybersecurity hardening. Conclusion: The integration of spatial and predictive analytics provides a robust and vital approach to managing ASF outbreaks. The system proves capable of significantly improving disease monitoring and forecasting, despite the need for specific technical refinements. Contribution: This research contributes a practical, data-driven framework for enhancing the resilience of the swine industry against ASF. Furthermore, it establishes a technological foundation that prompts and guides future national-level data integration and the expansion of AI capabilities in agricultural epidemiology.