A prediction method of urban water pollution based on improved BP neural network

Abstract

The existing methods for urban water pollution prediction have some problems, such as large prediction error and inconsistency with the actual pollution situation. A new urban water pollution prediction method is proposed. The water pollution data collection system of mobile GIS is used to collect urban water pollution data, analyse the overall structure of the water pollution data collection system, and classify the obtained urban water pollution data at different levels. The application concept of BP neural network is clarified, and the obtained urban water pollution data is entered into the network to obtain the urban water pollution prediction results. Genetic algorithm is used to improve the weights and thresholds obtained above, and the urban water pollution prediction model is constructed, and the prediction results of urban water pollution are output. Through the effective experimental analysis, it is concluded that the minimum error value is about 0.1%, and the prediction time is consistent with the actual time consumption.