应用ARMA模型超长期预测棉铃虫发生程度  

Superlong-Term Prediction of the Occurrence Level of Helicoverpa armigera by ARMA model

作　　者：黄荣华[1] 叶正襄[1] 汤建国 

机构地区：[1]江西省农业科学院植物保护研究所 [2]江西省彭泽县植物保护站

出　　处：《江西农业学报》1997年第1期40-44,共5页Acta Agriculturae Jiangxi

基　　金：江西省青年科学基金

摘　　要：根据赣北棉区１９５０～１９９１年棉铃虫发生程度的历史资料，应用时间序列分析方法，建立了自回归滑动平均（ＡＲＭＡ）预测模型。Ｗｔ＝－０．３６５５Ｗｔ－１＋０．２９２６Ｗｔ－２＋０．０３４１Ｗｔ－３＋εｔ＋０．８２０３εｔ－１＋０．３２６７εｔ－２＋０．３１４５εｔ－３该模型１９５３～１９９１年的历史符合率为８７．１８％，１９９２～１９９６年的预报准确率达１００％。On the basis of the data of historical occurrence of Helicovrepa armigera in the cotton-growing area of northern Jiangxi during 1950-1991, the following autoregressive moving average (ARMA) forecast model was established by the time series analysis method: W t=-0.3655W t-1 +0.2926W t-2 +0.0341W t-3 +ε t+0.8203ε t-1 +0.3267ε t-2 +0.3145ε t-3 Its historical coincidence rate in the period from 1953 to 1991 was 87.18%, and the forecast accuracy in the period from 1992 to 1996 reached 100%. According to the established model, the occurrence level of Helicoverpa armigera in 1997 and 1998 in northern Jiangxi was forcast.

关 键 词：ARMA模型 棉铃虫 发生程度 超长期预测 

分 类 号：S435.622.3[农业科学—农业昆虫与害虫防治]