Improved Grey Model Base on Exponential Smoothing for River Water Pollution Prediction

The aim of this project is to develop a river water pollution predictor. We present an improved Grey-based prediction algorithm to forecast the trend of the river water pollution. We adopted grey prediction as a forecasting means because of its fast calculation with as few as four data inputs needed. However, our preliminary study shows that the general Grey model, GM (1, 1) is inadequate to handle a volatile system. The general GM (1, 1) prediction generates the dilemmas of dissipation and overshoots. In this study, the prediction is improved significantly by applying the exponential smoothing technology and double accumulated generating operation (2-AGO). Based on exponential smoothing method, a new grey prediction model (ES-GM (1, 1)) was put forward, and it is applied to forecast the major pollutant of water quality of Yangtze River in Nanjing extension in the future five years. The forecast results show that the CODMN and BOD5 consistency are rising gradually every year, if no measures are adopted, the CODMN and BOD5 consistency will rise to 1.791 mg/L and 2.043 mg/L in 2012 respectively. The example shows that the prediction accuracy has been improved quite a lot in comparison with general model and thus points a novel direction to a higher modeling procedure.