Support vector machine with chaotic genetic algorithms for annual runoff forecasting

Annual runoff forecasting plays an ever-increasing role in water resources management as engineers are required to make competent forecasts of natural runoff to meet the increasing needs for irrigation, safe potable water and managements, etc. Traditionally, time series analysis is used for building mathematical models to generate hydrologic records in hydrology. Recently, Artificial intelligence system approaches such neural networks and support vector machine have been used successfully for hydrological time series modeling. However, the major shortcoming of neural networks is that the knowledge contained in the trained networks if difficult to interpret. Unlike neural network models, support vector machine (SVM) is based on the empirical risk minimization principle, and applies the structural risk minimization principle to minimize an upper bound of the generalization error, rather than minimizing the training error. This investigation presents a SVM model with chaotic genetic algorithm (CGA) as a promising method for hydrological prediction. The experimental results reveal that the SVM model with chaotic genetic algorithms results in satisfying predictions.