Disaster Risk Assessment System Using Support Vector Machine Algorithm and Risk Indicators

This study developed a disaster risk assessment system that integrates the Support Vector Machine (SVM) algorithm and risk indicators derived from social media data and official sources. Employing the Rational Unified Process (RUP) for system development and a mixed-methods design for evaluation, the study was conducted in the disaster-prone province of Sorsogon, Philippines. The system utilized a linear kernel SVM classifier to categorize social media posts as disaster-related or not and computed a Disaster Risk Index (DRI) using five key indicators: hazard, exposure, vulnerability, and capacity (hard and soft countermeasures). System features included data extraction modules, GIS-based visualization, and a security layer employing SHA-256 encryption. Usability testing using the USE questionnaire and qualitative interviews showed high levels of perceived usefulness, ease of use, and user satisfaction among disaster management personnel. The results identified Juban, Sta. Magdalena and Bulan as the municipalities with the highest disaster risk levels. The study concludes that the proposed system is an effective tool for enhancing disaster preparedness and recommends its future deployment with real-time data integration and expanded geographic coverage.