用带周期分量的逐步回归模型预测冬小麦产量的试验

本文首先对粗选因子进行主分量分析,得到立分量组合因子,将其连同预报量和主分量因子的周期分量一起嵌入逐步回归方程,来组建北方主要产麦区冬小麦产量预报模式,收到较好的效果。

An approach to winter wheat yield prediction using a model with orthogonal factors and periodic components

Since the prediction model is influenced by the multicolinearity of the selected factors and the unstability of the correlated coefficient, principal component analysis is used in this paper. Along with the compound factors of pricipal components of primary selected factors and the periodic components of prediction factors as well as the principal components itself, a stepwite regression model for predicting winter wheat yield in the major winter wheat areas of North Chian is established. Both its fitness and feasibility are better than those of the ordinary stepvvise regression models.