Water quality prediction model based on GRA and CNN-GRU

Water quality prediction plays a very important role in the prevention and control of water pollution. As the river water quality factors are time-series, uncertainty and correlation, in order to improve the prediction accuracy of water quality prediction model, a combined water quality prediction model based on gray correlation algorithm (GRA) and CNN-GRU combined model is proposed. Four common water quality indicators of chemical oxygen demand, dissolved oxygen, total phosphorus and total nitrogen in the water quality of Taihu Lake basin were selected for data prediction, and benchmark models such as GRU and LSTM were compared. The experimental results show that the predicted values of the combined GRA-CNN-GRU model fit better with the actual values, and the RMSE, MAE and MAPE errors are reduced compared with other models, and the model prediction effect is better