Application of chaos and fractal models to water quality time series prediction

This paper established a global chaos model to predict water quality time series that comprised a small amount of data, and the reducing-dimension chaos method was put forward to improve calculation of the global chaos method. Furthermore, sectioned variable-dimension fractal was also used to simulate and predict the same water quality time series. To investigate the prediction ability of the two chaos methods and fractal method, the auto regression (AR) model and grey model were used. The results illustrated that prediction accuracy of the global chaos method and the reducing-dimension method was higher than that of the AR model and grey model. The prediction result of the reducing-dimension method was the most satisfactory among the five methods. The results also indicated that the prediction accuracy of sectioned variable-dimension fractal method was better than that of the AR and grey methods. [Copyright &y& Elsevier]