论文部分内容阅读
对影响1982~1991年山东省及各类型区小麦病虫发生量(预报量y)的环境因素(预报因子XJ),在一元线性回归初选的基础上,经逐步回归、逐步判别、多元回归和多级判别分析进行精选,共建立全省和各类型区预测模型186个,从中筛选出最优预测模型30个,于1993年和1994年应用,平均预测准确率达90%以上。
On the basis of linear regression, the environmental factors (forecast factor XJ) affecting the occurrence of wheat diseases and insect pests in Shandong Province and various types of areas from 1982 to 1991 were estimated by stepwise regression and stepwise regression. Multiple regression And multi-level discriminant analysis. A total of 186 prediction models in the whole province and various types of regions were established. 30 optimal prediction models were selected and applied in 1993 and 1994 with an average prediction accuracy of over 90%.