Prediction model of chaotic time series based on support vector machine and its application to runoff

Chaos theory and support vector machine have great capability of dealing with nonlinear matter.Based on the phase-space reconstitution theory,the prediction rmdel of chaos time series is built by using the support vector machine in this paper,the method,the characteristic,and the selecting of the key parameters in the modeling is discussed.Fnrstly the phasespace re-constitution is made by saturated correlation dimension,so that information of monthly runoff series is profoundly investigated.At the same time,the maximum Lyapunov exponent is computed using the improved small-data method,and it is used to recognize the chaotic feature of the monthly runoff at YiChang.In the application of chaos time series using support vector machine model to predict the rmnthly runoff,the RBF kernel function is introduced,which simplified the course of solving non-linear problems.It is shown by the study case that the rmdel proposed in the paper can process a complex hydrological data sieres better,and has better generalization and prediction accuracy.