影响流域土壤流失的降水量的PPAR(K)预测模型

MODEL OF PREDICTION OF PRECIPITATION RESULT IN SOIL CROSION USING PPAR(K)

介绍了投影寻踪自回归（ＰＰＡＲ（Ｋ））的基本思想及其算法。用ＰＰＡＲ（Ｋ）和均生函数（ＭＧＦ）两种方法对泸州市和北碚市降水量的时间序列进行了建模预测。结果表明ＰＰＡＲ（Ｋ）模型具有较高的稳定性和预测精度。

The main idea and realization algorithm of the projection pursuit regression based on SMART are introduced.Comparisons of precipitation prediction between PPAR(K) and mean generating function methods are carried out for cities Luzhou and Beibei.The results show that the precision of prediction using PPAR(K) model is much better than that of mean generating function model.