Information Completeness Assessment of Demand Posts on Social Media for Enhanced Disaster Response  
Author Cancan Zhang

 
Co-Author(s) Yudi Chen

 
Abstract Nowadays social media platforms have become an important channel for disaster victims to seek assistance, share disaster information, and organize self-help and mutual aid efforts. In order to enhance disaster response efforts, this paper assesses the information integrity of demand posts on social media and suggests improvements for decision makers and disaster victims. Based on the demand-related information shared on social media, this study uses a fine-tuned large-scale language model (f-LLM) to identify specific demand posts, categorize them by urgency level, and analyze the correlations between urgency and the completeness of the information. The study found a weighted correlation coefficient of -0.22 between demand hierarchy and information completeness, with a confidence level of <0.001, indicating a statistically significant negative correlation. This suggests that as demand hierarchy increases, information completeness tends to decrease. To this end, we have made relevant recommendations for improvement.

 
Keywords Social media, Disaster emergency management, f-LLM
   
    Article #:  DSBFI25-72
 
Proceedings of 3rd ISSAT International Conference on Data Science in Business, Finance and Industry
January 6-8, 2025 - Da Nang, Vietnam