A Machine Learning Framework for Fire Risk Prediction With Response and Proximity Insights

Advances in artificial intelligence (AI) and machine learning (ML) have significantly enhanced fire risk assessment by enabling predictive analytics, real-time decision support, and optimized emergency response. Accurate fire risk assessments are crucial for prioritizing high-risk zones and optimizing resource deployment to minimize damage and enhance safety. In this work, we introduce novel fire risk models and propose a comprehensive ML-based framework for fire risk prediction that supports data-driven decision-making for fire and emergency response services. Our models incorporate response performance and service proximity to assess the impact of incidents more effectively within a city. The proposed framework provides an end-to-end ML pipeline that integrates diverse data sources to construct a dataset, compute risk scores, analyze key features, and formulate fire risk prediction as a regression problem. Additionally, it evaluates multiple regression models to analyze risk variations at both the incident and neighborhood levels. Experimental results demonstrate that our proposed models achieve a high degree of alignment between predicted and actual risk scores with minimal error. This framework captures valuable spatial risk patterns and can be used as a reliable tool for fire risk assessment, resource allocation, response strategy improvement, and urban safety planning.