From Latency, Through Outbreak, to Decline: Detecting Different States of Emergency Events Using Web Resources.
Zheng Xu,Xiangfeng Luo,Yunhuai Liu,Kim-Kwang Raymond Choo,Vijayan Sugumaran,Neil Yen,Lin Mei,Chuanping Hu
DOI: https://doi.org/10.1109/tbdata.2016.2599935
2018-01-01
IEEE Transactions on Big Data
Abstract:An emergency event is a sudden, urgent, usually unexpected incident or occurrence that requires an immediate reaction or assistance for emergency situations, which plays an increasingly important role in the global economy and in our daily lives. Recently, the web is becoming an important event information provider and repository due to its real-time, open, and dynamic features. In this paper, web resources based states detecting algorithm of an event is developed in order to let the people know of an emergency event clearly and help the social group or government process the emergency events effectively. The relationship between web and emergency events is first introduced, which is the foundation of using web resources to detect the state of emergency events imaged on the web. Second, five temporal features of emergency events are developed to provide the basis for state detection. Moreover, the outbreak power and the fluctuation power are presented to integrate the above temporal features for measuring the different states of an emergency event. Using these two powers, an automatic state detecting algorithm for emergency events is proposed. In addition, heuristic rules for detecting the states of emergency event on the web are discussed. Our evaluations using real-world data sets demonstrate the utility of the proposed algorithm, in terms of performance and effectiveness in the analysis of emergency events.