火灾预测的改进GM(1,1)模型

Application of improved GM(1,1) models in fire forecasts

火灾的发生受很多因素的影响,有的因素已知、有的未知,符合灰色系统的特征,从而可以利用灰色理论进行预测.现从传统GM(1,1)预测模型构造原理出发分析其存在的理论缺陷,通过采用平均斜率法、等维替换和变换灰色预测公式的方法建立改进的GM(1,1)模型,以适应火灾频数序列的波动特性,从而达到精确预测的目的.我国1997—2006年火灾数据预测计算表明,新方法有满意的拟合和预测效果,从而为火灾预测精度提高提供了新的途径.

There are many factors for fire＇s occurrence. These factors accord with the characteristics of a grey system and thus the fire could be predicted by the grey theory. An improved GM（1, 1） model is put forward according to the principle of the tradtional GM （1, 1） model and its defect. It is fit for the characteristic of the fire frequence list. Fire data of the years 1997-2006 in our country were used. It shows that the new model has satisfactoried fitting effects and thus creates a novel direction to higher modeling procedures for fire forecasts.