Water Quality Prediction Method Based on LSTM-BP

Abstract

Water quality prediction is of practical significance not only for the planning, evaluation, and management of the water environment, but also for the prevention and control of water pollution. In order to improve the accuracy of water quality prediction, an LSTM-BP combined model algorithm based on Long Short Term Memory Neural Network (LSTM NN) and BP neural network is proposed. Taking the water temperature data of No.6 large-scale integrated observation buoy on the Yangtze estuary as an example, a time series prediction model framework is established, and the data processing to model simulation is completed with the help of Python to realize the water quality prediction based on LSTMBP. The method is compared with LSTM model and BP model, the experimental results show that the time series predicted by LSTM-BP is more accurate. This LSTM-BP model can be effectively applied to the prediction of water quality indicators and the early warning and prediction system of water quality trends.