Precipitation Time Series Predicting of the Chaotic Characters Using Support Vector Machines

Based on the powerful nonlinear mapping ability of support vector machines, the predicting model of support vector machines in combination with takens´ delay coordinate phase reconstruction of chaotic time series has been established. Yearly precipitation time series is of the chaotic characters, thus this model is used to try predicting the precipitation. Because of the peculiarity of precipitation time series, the mean-square-error is used as the criterion to choose the embedding dimension and model parameters. Case study proved the precise of this model to predicting the precipitation. Besides, this result also shows that support vector machines is one of tools to study chaotic time series.