Water quality prediction research based on SSA-LSTM model

Abstract

In recent years, due to too much attention to economic development and ignoring the ecological environment, environmental pollution and resource shortage have become increasingly serious, and the protection of water resources is the most important part of ecological protection. Therefore, by establishing relevant prediction models, it is particularly important to predict the trend of various indicators of water bodies in the future based on the data that has been monitored. This paper studies water quality prediction models, aiming to infer the trend of water quality data by mining and analyzing the regularity information hidden behind water quality data. Firstly, a sparrow search algorithm is established to optimize the water quality prediction model of generalized Long Short-Term Memory network (LSTM). Using the global optimization characteristics of the Sparrow Search Optimization Algorithm (SSA), the key parameters can be optimized, and the high-precision approximation ability of the LSTM is combined to establish a water quality prediction model. The experimental results show that the root mean square error of the water quality prediction model of the SSA to optimize the LSTM is 0.0039543 in the training of PH value, which not only has a more stable prediction curve compared with the classical water quality prediction model, but also shows better prediction accuracy in the overall prediction of PH water quality index.