基于小波ANFIS的诸暨站降雨量预测模型

Rainfall Prediction Model of Zhuji Station Based on Wavelet ANFIS

【目的】建立适合诸暨站的降雨量预测的模型。【方法】采用小波分析和ANFIS相结合的模型预测方法,首先利用小波分析的多尺度分解功能,将诸暨站降雨量进行分解,然后利用ANFIS的自适应非线性逼近功能对分解后的序列进行预测,最后利用小波重构对预测的分解结果进行重构。【结果】小波ANFIS模型对诸暨降雨量的预测值与原始值相对误差在30%以内,精度达到要求。【结论】小波ANFIS模型可作为诸暨站的降雨量预测模型。

【Objective】The purpose of this study is to establish a better rainfall prediction model suitable for Zhuji station.【Method】This paper used the combination of wavelet analysis and ANFIS model prediction method.Firstly,the rainfall of Zhuji station was decomposed by the multi-scale decomposition function of wavelet analysis,then the decomposed sequence was predicted by the adaptive nonlinear approximation function of ANFIS,and finally the decomposition result was reconstructed by wavelet reconstruction.【Result】The prediction accuracy of the wavelet ANFIS model for Zhuji rainfall was less than 30%,which could meet the requirements.【Conclusion】The wavelet ANFIS model can be used as the rainfall prediction model of Zhuji station.