Research On Modeling Of River Water Quality Prediction Based On RF-IFA-GRU

The river system is an ecological environment system full of unstable factors and complex and changeable.The river water quality prediction and early warning is helpful to water environment management and ecological protection.With the help of mathematical models,data analysis and computer science and technology,the characteristics of natural river water bodies are studied and analyzed.Water quality evaluation and the establishment of a non-mechanical water quality prediction model based on receiving water bodies can provide a scientific basis for the formulation of emergency treatment measures for sudden river water pollution incidents,scientific formulation of water environment planning,and comprehensive water pollution prevention and control.In order to explore new water quality prediction methods,intelligent calculations are used for water quality prediction modeling,through the analysis and screening of water quality characteristics in natural rivers,a computer-assisted mathematical model with neural network as the core is established,in order to improve the river system Water quality forecast and early warning and comprehensive treatment provide scientific basis.It was discussed a combination model of intelligent forecasting based on Random Forest(RF),Improved Firefly Algorithm(IFA)and Gated Gated Unit(GRU).Main research work and results of this paper are as follows:1)Using Dissolved Oxygen(DO),a key parameter of river water quality,as the water quality evaluation index,the water temperature,p H,and p H values collected by the high-frequency water quality monitoring station established by the United States Geological Survey(USGS)in the Potomac River are selected.Water quality parameters such as conductivity and turbidity are data.Random forest algorithm is used to select data features,reduce the interference of irrelevant data in the model prediction process,and reduce the workload of model calculation.2)Use Lévy flight to adjust the random item generation scheme of the Firefly algorithm,improve the ability to search for the optimal value of the population,and improve the adaptability of the algorithm;use the variable step size to replace the fixed step size factor to make the algorithm better in the later calculations To lock the optimal value.By improving the Firefly algorithm to improve the algorithm to solve the target problem,the convergence rate and solution accuracy.The simulation results show that the optimized and improved Firefly algorithm has better adaptability and accuracy.3)The RF algorithm,IFA algorithm and GRU algorithm are combined to apply to river water quality prediction modeling.The IFA algorithm is used to optimize the parameters of the GRU network.The data after feature selection is used for prediction experiments,and it is combined with RF-FA-GRU and RF.-Compare the prediction results of IFA-LSSVM.Experimental results show that the prediction accuracy of RF-FA-GRU is higher.