将灰色微分方程与自记忆函数相结合构建灰色自记忆模型,并对杭州市年降水量序列进行拟合和预测,采用逐步回归周期分析法提取该序列的周期项后,再对残差序列建立灰色自记忆模型。结果表明,采用残差序列建立的模型,模拟和预测精度大幅提高。 The gray self-memory model is constructed by combining the gray differential equation with self-memory function. The annual precipitation series in Hangzhou are fitted and predicted. The periodical term of the sequence is extracted by stepwise regression period analysis, and then the residual sequence is established Gray self-memory model. The results show that using the model established by the residual sequence, the accuracy of simulation and prediction is greatly improved.