Prediction of PM2.5 Concentration Based on Bayesian Optimization Random Forest

Nowadays, people pay more and more attention to air quality, and smog as a serious air pollution phenomenon has received greater attention. As the main culprit of haze phenomenon and an important indicator of air quality, it is of great significance to forecast PM2.5. Aiming at the prediction of PM2.5 concentration, this paper proposes a method of using random forest algorithm to build a regression prediction model and then predict PM2.5 concentration. This method mainly uses meteorological factors related to PM2.5 (such as dew point, temperature, humidity, etc.) and historical data of PM2.5 concentration in the past three hours as input to build the model. Forecast future hourly PM2.5 concentrations. In this paper, the random forest model is successfully constructed, and the Bayesian algorithm is used to optimize the random forest model, and the PM2.5 concentration is well predicted. At the same time, this paper also constructs the decision tree, support vector machine, intelligent gray Wolf algorithm optimized random forest model and Bayes algorithm optimized random forest model to compare. The superiority of Bayesian algorithm in optimizing stochastic forest model is proved.