大袋蛾发生级别的两种回归测报模型研究

Research on two kinds of regression forecast models of Ciania variegata generation class

采用双重筛选逐步回归和主分量旋转回归两种方法 ,对影响大袋蛾发生率、有虫株率、虫口密度的气象因子进行分析 ,剔除不显著影响因子 ,保留显著因子 ,建立 9个测报模型 在此基础上 ,采用加权平均法将因变量 3个指标综合为一个指标 ,并划分等级 利用已建立模型进行回测 ,历史符合率达 80

Acouple of methods were adopted of stepwise regression of doubing screening and rotatory regression of principal component to analyse the meteorologic factors affecting the frequency of occurence of ciania Variegata,having worm strain rate,and population density insignificant factors were omitted,significant factors were emitted, significant factaor were resered and 9 forecasting models were established.On this basis the method of weighed mean mas adopted to integnate the 3 indicatrix,into 1,and the nanks were Compartmentalized.The estabished models,were used to return to measure and the historical coincidence rate come to the above of 80 percent.