中文摘要：

在对河南登封地区火灾发生状况进行有效分析的基础上,以林火发生与其紧密相关的气象因子为基础,运用BP人工神经网络建立了林火发生预测模型,分析得出最好的网络结构为6∶5∶1,建模精度为0.97,98%的样本拟合误差绝对值不超过0.000 7,取得了较好的拟合效果。

英文摘要：

Meteorological factors related to forest fire are temperature,humidity,precipitation,and wind.Based on the analysis of the specific situations of forest fire occurrence in Dengfeng County,Henan Province,China,the prediction model of forest fire was established by using BP artificial neural network.It was found that the preferable neural network was 6∶5∶1,the fitting degree of accuracy was 97%,and 98% of the samples＇ fitting errors were not more than 0.000 7,indicating that the fitting was satisfactory.