Prediction of Water Quality Monitoring Indicators Based on Random Forest

Abstract

Water resources are not only the material basis to support the current social and economic development, but also an important resource to maintain people's daily life. Therefore, the protection of water resources from being infringed not only requires the cooperation of relevant researchers, but also is closely related to every resident in the society. Everyone needs to pay more attention to it. In the current society, the water environment in different regions is generally protected through multiple processes. Among them, the first line of defense is the monitoring of water quality and the prediction model of indicators. The monitoring and early warning model of water quality indicators is generally based on the real-time monitoring of the content of various substances in the water environment, so as to comprehensively evaluate the water environment and predict the water quality in the future period according to the concentration of different substances. The existing water quality indicator monitoring and prediction model not only is an important technology to control water environmental pollution in the current society, but also can reflect the quality status in the target waters and the future water quality development trend in all aspects. More professional water quality related data can be collected through the existing monitoring models of various indicators of water quality. Based on the in-depth analysis of these data, the main pollutants in the target waters, the current pollution situation and the possible future damage can be fully understood. On the other hand, this water quality indicator monitoring model has also provided assistance for professionals to develop more scientific and reasonable water pollution control plans. In this paper, the existing water quality index monitoring and prediction model was updated by using the stochastic forest algorithm model. The stochastic forest algorithm model generally predicts the indicators and pollution status of the target water area in the future by processing the relevant data of the water quality indicators of the target water area. Finally, the performance of this new water quality monitoring index and prediction model has been improved by about 26% on average.