Scope Detection And Evolution Process Analysis Of Disaster Based On Social Media

Accurate and timely extraction of emergency information,identification of disaster-stricken areas and situation analysis constitute the cornerstone of emergency response scenarios.With the development of mobile communications technology and smart devices,social media rapidly spread in almost every aspect of modern life.The spontaneous,immediate and extensive social media data has been widely used in disaster emergency response.The aim of this thesis is to use the data collected through social media platform to detect disaster-stricken areas timely and track the development trend of disaster events.For this purpose,this study takes two earthquake disasters and a rainstorm disaster as examples to extract the spatiotemporal features and semantic information of social media data.The main work of this paper is as follows:(1)Rapid perception of disaster-stricken areas: this paper constructs an analogy between changes in the biological population size over time and changes in citizensensor data size over distance,thereby proposing a Spatial Logistic Growth Model(SLGM)to describe the spatial growth of citizen-sensor data after an earthquake,which reveals the mapping mechanism between the spatial distribution of citizen sensing signals and disaster-stricken areas;a framework is then developed to estimate the earthquake impact area by combining social media data and other auxiliary data based on the SLGM.To demonstrate how our approach leverages citizen-sensor data from social media to quickly locate earthquake impact areas,we used representative case studies,namely,the 2015 Nepal earthquake and the 2017 Jiuzhaigou earthquake.The reliability of our approach is demonstrated in these two earthquake cases by comparing the detected areas with official intensity maps,and the time sensitivity of the social media data in the SLGM is discussed.(2)Disaster situation perception: A method of extracting and expressing the disaster-related spatio-temporal semantic knowledge contained in social media data is proposed.Firstly,the feature words are extracted and the word relational network is constructed.Secondly,the Louvain community detection algorithm is used to classify the network for mining the text topic.Finally,a knowledge representation scheme and reasonable storage method of the information extracted before is designed based on knowledge map.A disaster social media knowledge network is constructed,which integrates semantic and temporal information.This study takes the 2012 Beijing rainstorm event as an example,using the proposed method to analyze the development of this disaster.Experiments show that when disasters happen,people can quickly extract valuable emergency information to detect disaster impact areas and understand the emergency situation more in-depth through analyzing the social media sensor data.The proposed earthquake impact area detected method solves the problem that traditional methods are constrained by the amount of social media data and the long response time.At the same time,the method of building social media disaster knowledge network can help us to enhance our situational awareness and improve our understanding and response to such events.