Application of a Hybrid ARIMA and Neural Network Model to Water Quality Time Series Forecasting

Abstract

In this paper the water quality forecasting at the Nanjinguan water quality monitoring station of Yangtze River, China, is presented. The time series data used are weekly water quality data obtained directly from Nanjinguan station measurements over the course of five years. In order to forecast water quality, hybrid models consisting of Autoregressive Integrated Moving Average (ARIMA) models and Artificial Neural Network (ANNs) models were developed. The ARIMA models were first used to do water quality forecasting of the time series data and then with the obtained errors ANNs were built taking into account the nonlinear patterns that the ARIMA technique could not capture, in order to reduce potential errors. Once the hybrid models were developed 38 samples out of the data for the station were used to do the water quality forecasting and the results were compared with the ARIMA and the ANNs models worked separately. Statistical error measures such as the root mean square error (RMSE), the mean absolute percentage error (MAPE) and correlation coefficient (R) were calculated to compare the three methods. The results showed that the hybrid models predict the water quality with a higher accuracy than the ARIMA and ANNs models in the examined station.