摘要
为了研究某日某林业局能否发生人为火，用大兴安岭地区１９７３～１９８１年计１５年间的资料建立Ｌｏｇｉｓｔｉｃ回归模型；并对预报因子进行了似然比检验，剔除了不必要的因子。采用Ｌｏｇｉｓｔｉｃ回归模型算法简便，容易在计算上实现。着火样本和不着火样本回报判对率分别为７９．７３％、８６．４４％；预报判对率１９８８年分别为７７．１０％、８０．３３％；１９８９年分别为７６．８８％、７８．５４％。
In order to forecast forest fire occurrence.the Logistic regression models were devdoped using data collected from DaXing'anling region from 1973 to 1987.The likelihood ratio examination for forecast factors was done by scratching out unimportant factors.It is very easy to calculate and realize on computer using Logistic regression models.The correct-rate of forecast backward of fire and no -fire samples are 79.31%, 86.44% and forecast forward are 77.10%、 80.33% for 1988, 76. 88%, 78. 54% for 1989.
出处
《东北林业大学学报》
CAS
CSCD
北大核心
1995年第1期102-107,共6页
Journal of Northeast Forestry University
基金
黑龙江省防火基金
