冬小麦生产经济收益的多元正交多项式回归预测模型

Multivariate Orthogonal Polynomial Regression Forecasting Model of Economical Effect on Winter Wheat Production

对冬小麦灌水，施P2O5、施纯氮各三水平条件下，进行全因素试验，得到了高产高收益的三因素的最佳组合。采用多元正交多项式回归方法建立了冬小麦生产经济收益预测的多元正交多项式回归模型。经F检验，说明施磷对经济收益影响显著，灌水对经济收益影响亦显著，而加纯氮影响不显著。方差估计σ＝435．7949，预报每公顷经济收益W的95％置信区间为W±1．96σ＝W±854．158，效果较好。

We have got highly yielding and highly economic effect combinations through complement factor experiment for watering and fertilizing with three levels for wheat production. We have build the multivariate orthogonal polynomial regression forecasting model of the economical effect on winter wheat production. The factors of P2O5 and watering are significant at level α=0. 01 for the economical effect by F test, but nitrogen application is not significant. Variance estimate is σ =435. 7949, the confidence interval of economical effect estimate value W in 95% is W ±1. 96 σ = W ±854. 158 per ha.