产量变化的主要成份分析

利用历史产量序列重构二维时变矩阵进行自然正交分解，得到了3种影响产量变化的主要成份。通过分析发现，这些主要成份不仅包括能反映生产力变化的趋势分量和能表达气候影响效果的短期振荡分量，还发现了一种由气候决定的趋势作用分量，对产量有着不可忽视的影响。这些分量相互独立，无预先设定的函数模型，为产量分析和预报提供了较为客观和完备的历史变动序列，对提高产量预报精度有着重要意义。

The Main-elements Analysis of Yield Variation

In this article, a two-dimension yield matrix reformated from a historic yield sequence and then isolated by the way of spontaneous orthogonal division, three main-elements which have influences on yield variation were obtained. These elenments including not only a tendentious sequence expressing artificial effect and anogher sequence expressing wether’s effect, but also a action besides man and wether which distinctly affects crop’s product. These elements are independent from each other and have no previous f-models, so by this way, physical and perfect sequences can be supplyed for the product analysis or prediction. This result is important for the improvement of the precision of product prediction.