二代玉米螟发生动态的判别分析数学模型及其应用

Mathematical Models of Discriminant Analysis for Forecasting the Occurrence Dynamics of the Corn Borer and Its Application

本文以山东省胶南市１９８５～１９９４年二代玉米螟虫情与相应年份的气象资料为研究对象，通过逐步回归分析，从１４个因子中筛选出３个主要预测因子，应用基于ＢＡＹＥＳ准则下的判别分析原理，建立了３个级别的判别分析数学模型。经回报验证，历史拟合率为９０００％，将１９９５、１９９６年作为独立样本进行试报，结果与实际一致。

Based on the historic data of 10 years from 1985 to 1994 in Jiaonan,Shandong Province,and through the method of progressively regression analysis,3 main factors were selected from 14 elements,By applying principle of discriminant analysis,3 mathematical models were set up for forecasting the occurrence dynamics of the second generation of the corn borer.The results tested the pest data showed that the fitting rate was 90.00%,and that accuracy of 1995 and 1996 forecast was right.