摘要
本文论述了Ｆｉｓｈｅｒ准则下多级判别的基本原理，并应用该原理将山东省胶南市第二玉米螟百株累计卵量的３个影响因素压缩为一个新指标（线性判别函数），用这一个新指标对样品按距离最小准则判别归类。结果表明，历史拟合率为１００％，两年预报，其结果与实际一致。
In this paper,the basic principle of discriminant analysis by Fisher's rule is analysed,a new index (linear discriminant function)is set up on the basis of the three influencing factors for the data about 100 plant egg amount of the second generation of the corn borer in 10 years of 1985～1994 in Jiaonian City,Shandong Province. The tested results of the pest data by using the index show that the fitting rate is 100% for forecasting of other the population dynamics of pests.
出处
《运筹与管理》
CSCD
1998年第4期67-71,共5页
Operations Research and Management Science
