G. H. Li, S. Lu, P. Mei, H. P. Zhang and S. M. Lo (2014) Influences of Time, Location, and Cause Factors on the Probability of Fire Loss in China: A Correspondence Analysis. Journal/Fire Technology 50 1181-1200. [In English]
Web link: http://dx.doi.org/10.1007/s10694-014-0385-6
Keywords: Fire statistics; Fire loss; Correspondence analysis; Influencing factors; STATISTICS

Abstract: A novel application of correspondence analysis for identifying the relationships between fire loss and time, causes, and locations is proposed. The fire loss is divided into four levels which are [1, 1 k], [1 k, 10 k], [10 k, 100 k], and [100 k, a] Yuan (RMB) (1YUAN = 0.1647US, $), where 1 k equals to 1,000 Yuan. The results indicate that the loss levels have significantly related to months, time of day (hour), years, fire causes, and locations. Fires occurring in January, February, and June tend to be strongly associated with the loss level of [1, 1 k], while the loss level of [1 k, 10 k] is more likely to occur in May and August. It shows that September and December have strong relationship with the high loss level of [100 k, a]. Regarding the time of day, low loss level of [1, 1 k] is associated with the period from 10 a.m. to 10 p.m.; while high loss levels tend to occur between 2 a.m. and 6 a.m., the level of [1, 1 k] tends to be related to the causes of smoking, playing with fire, and others, while the level of [1 k, 10 k] is mainly caused by the electrical fire and spontaneous combustion. The locations of traffic, production operation places and commerce buildings, have been significantly related to high loss level. This paper introduces the application of correspondence analysis to fire losses, demonstrating the possibility of using CA to identify the relationships between fire losses and influencing factors. The achieved information can improve loss assessment of fires at a given time or location.