论文部分内容阅读
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 study whether a certain forestry bureau of a certain day could be man-made fire, a logistic regression model was established based on the data from 1973 to 1981 in the Greater Xing’an Mountains, and the likelihood ratio test of the forecasting factors was carried out to eliminate unnecessary factors. Logistic regression model using the algorithm is simple and easy to calculate. The rates of judgment on fire samples and non-fire samples were 79.73% and 86.44%, respectively; the forecast judgment rates were 77.10% and 80.33% respectively in 1988 and 76.88% and 78% respectively in 1989 .54%.