旱涝趋势的投影寻踪预测模型

MODEL FOR DROUGHT AND FLOOD TREND PREDICTION USING PROJECTION PURSUIT REGRESSION

介绍了投影寻踪回归（ＰＰＲ）建模的基本原理和算法实现。选用太阳黑子年平均数和旱涝分型的转移概率作因子，建立了长江中下游旱涝趋势的ＰＰＲ预测模型，并与用相同资料建立的Ｂ－Ｐ神经网络模型预测结果进行了比较。

A fundamental idea and implement algorithm of projection pursuit regression (PPR) are introduced.The trend prediction model for drought and flood in the middle and lower reaches of the Yangtze River is developed with the average numbers of sunspots yearly and the transfer probability of drought and flood types as the factors.The comparisons of the prediction results batween the PPR model and the B P neural network model are performed using the same data.