Research And System Implementation Of Air Quality Forecasting Based On Ensemble Learning

Air pollution problems in Sichuan Province have become frequent and gradually become one of the focal points of Internet public opinion,as a result,people are paying more and more attention to air quality,it is especially important to establish an air quality prediction system in a timely manner.In the treatment of air pollution problems,air quality prediction is the most important task,by timely forecasting air quality,it is convenient for people to reasonably arrange travel,and it is also convenient for the government to take effective measures.Therefore,the research of air quality prediction and the implementation of prediction system have practical significance.Based on the collected air quality related data,the relevant data was analyzed,the correlation between the pollutant index and the air quality index was explored,and the inherent relationship with air quality in different seasons and regions was explored.The key to air quality prediction is the accuracy of the model,by using the historical air quality data of Sichuan Province from 2015 to 2018 as a sample set,using random forests,gradient boosting trees,and XGBoost to build air quality models,in order to further improve the prediction accuracy,Bayesian optimization is used to tune the parameters of each model and fuse them.The effectiveness of Bayesian optimization is verified through comparative experiments,it is also found that the fusion model based on Bayesian optimization of random forest,gradient boosted tree,and XGBoost has better prediction accuracy than other models in air quality index prediction.At the same time,the air quality prediction system was designed and implemented using Python,Django and Echarts,and the fusion model of random forest,gradient boosted tree and XGBoost based on Bayesian optimization was applied to the system’s air quality prediction module.The air quality prediction system adopts the MVC design mode and includes three parts from the bottom up: the data layer,the logic layer,and the application layer.It mainly implements functions such as the collection agent pool,data collection,data visualization management,and air quality index prediction.By implementing and testing the function of the air quality prediction system,it is explained that the air quality prediction system basically meets the needs of users and preset goals.It can help the public understand the future air quality situation in advance,so as to make more reasonable travel arrangements,and it can also help government departments to take corresponding measures in time for the future air quality situation,and promote the healthy development of society.