MODELING THE SPREAD OF EMERGENCY ALERTS THROUGH SOCIAL AND SENSOR NETWORKS

Abstract

Timely dissemination of emergency alerts is critical for minimizing risk and enabling coordinated responses in urban environments. However, existing systems suffer from delayed propagation, fragmented communication, and limited integration between human and sensor networks. To address these challenges, this research presents a hybrid modeling framework to analyze and optimize the spread of emergency information across integrated social and sensor networks. The system combines Internet of Things (IoT) sensors, mobile devices, and human communication channels within a unified graph-based architecture to capture real-world alert propagation dynamics. The dataset consists of simulated and real-world emergency scenarios, such as fire incidents, traffic accidents, and environmental hazards, derived from sensor streams and social media logs. Data preprocessing includes noise removal, feature normalization, missing value imputation, and temporal synchronization. To enhance prediction and dissemination efficiency, deep learning models are incorporated, including Bald Eagle Search Optimized Spatio-Temporal Long Short-Term Memory (BESO-ST-LSTM) networks for modeling network interactions and temporal pattern analysis using Python 3.11. The model evaluates key performance parameters, including alert latency, propagation coverage, node influence, and reliability. Experimental results show that the proposed method improves emergency alertsachievingaprediction accuracy of 0.983%, a Loss of 0.052%, a Precision of 0.953%, a Recall of 0.54%, and an F1-Score of 0.954. Furthermore, adaptive routing and priority-based alerting enhance system robustness and scalability. The findings demonstrate that integrating deep learning with hybrid social-sensor networks enables faster, more reliable emergency communication, supporting the development of intelligent and resilient smart city infrastructures.