of Tsing Hua University

Beijing, China

of Tsing Hua University

Beijing, China
SEARCH FILTERS
Time filter
Source Type

Qiao J.,Capital University of Economics and Business | Li Y.,China Life Insurance Group Company | Li Y.,of Tsing Hua University
C e Ca | Year: 2017

The identification of high fire risk areas is one of the most important research directions in fire risk management. In order to develop a comprehensive solution to the problem, this paper presents a combined technique based on historical data of fire accidents within sub-districts in lieu of the behavioral or physical characteristics of specific places. Specifically, the fire probability is analyzed based on discrete-Time Markov chain (DTMC), which interprets fire state transitions of a sub-district. Then, high risk areas are prioritized by k-means clustering to avoid the bias of experts, and the k-means clustering model is solved with dynamic programming. Finally, a case study is performed on the 15 sub-districts in Xicheng District, Beijing, revealing that 5 of the sub-districts are under high fire risk in the next month, and that Guang'anmennei and the other 4 sub-districts tend to have high fire risk in the long run. The author suggests making the first 5 sub-districts the focal point of fire hazard identification by Xicheng Fire Brigade, and building more professional fire stations and safety communities to improve fire prevention.

Loading of Tsing Hua University collaborators
Loading of Tsing Hua University collaborators