Abstract
Natural disasters pose significant challenges to affected communities, governments, and relief organizations, necessitating innovative disaster response and recovery strategies. The rise of social media platforms in recent years has transformed disaster management, presenting both opportunities and complexities. This study delves into the multifaceted role of social media in shaping natural disaster responses. Researchers examine its utilization before, during, and after disasters for information dissemination, relief coordination, resource mobilization, and emotional support. Additionally, employing classification models like Support Vector Machine (SVM), Logistic Regression (LR), and Decision Tree (DT), the study assesses their performance using accuracy, recall, precision, and F1 score metrics. The SVM model achieves 94% accuracy, with 92% precision and 94% recall, resulting in a 95% F1 score. LR demonstrates similar performance, scoring 95% across accuracy, precision, and recall, yielding a corresponding 95% F1 score. In contrast, the DT model outperforms both, achieving 97% accuracy, 96% precision, and recall, culminating in an impressive 97% F1 score. These results highlight nuances in model efficacy, with DT showcasing superior performance. Moreover, the DT model exhibits a faster computation time at 37.203 ms compared to SVM and LR. This research sheds light on the dynamic relationship between social media and disaster response, offering insights for stakeholders to harness its potential in bolstering preparedness, response, and resilience during natural disasters.