灾变预测在火灾预测中的应用

简述了灰色理论中灾变预测的特点及其建模过程，并将其应用于湖南省和山东省火灾预测之中。经检验，模型的精度等级为好，模型的预测结果可靠，对于消防部门火灾的预测有一定帮助。

APPLICATION OF DISASTROUS FORECASTING IN PREDICTION OF FIRE

This present paper presents a brief introduction to the setting-up procedures for the calamity forecast model, a model built up in accordance with the gray theory with a summary of the particular features of the above model. As is known, calamity forecasting is mainly done based on the data that are expected and the data unexpected. Since some factors are certain but others uncertain in the fire, that is, partially known and partially unknown, the gray theory established by professor Deng Julong in the 1980s can be successfully applied in actual fire predication. In the process of fire prediction, chief factors involved are the abnormal values in the data, that is, whether the unexpected values are larger or smaller than the given threshold value, the ordered series of data formed by the times when the values appear. A new series of values can be obtained from the mapping between the abnormal values to the values resulted when the calamity takes place. And then the model GM (1,1) is utilized to make the predication. The predicative result will eventually be the difference between the time when the next or the next few abnormal values appear and the last time when the model is proposed. The precision of the model has so far been tested and proved to be efficient by employing the method of posterior difference checking. Since the model is made based on the recent data, the precision of the predication for near future will be higher than that to be made for the far future. The present model has been tried a few times to the fire forecasting in provinces of Hunan and Shandong with the results showing the accuracy in function, reliable and practical in the prediction. As the results of the fire predication is made when fire frequently broke out, and therefore, it is of a certain guiding significance to the fire department in fire predication.