Runoff Forecast Based on Wavelet Analysis and PSO Optimized ANFIS' Combined Model
Fan Li · China Rural Water and Hydropower · 2011
Affected by a variety of factors,runoff series are charactered by very complex change characteristics.If conventional methods are used to predict them directly,their accuracy is seldom very high.If we decompose the runoff series into more simple series,we will use nonlinear prediction methods to predict it and its accuracy will be improved.In this paper,wavelet analytical multi-resolution decomposition function is used to decompose the runoff sequences,it can reduce runoff series' complexity factor.PSO is used to optimize the network parameters of adaptive neuro-fuzzy inference system.It can improve the precision determination of structure parameters.It uses wavelet analysis and PSO optimization ANFIS combination model to forecast runoffs.The example shows that the model can improve the accuracy of runoff prediction,the prediction result is better.