Air Quality Analysis And Time Series Prediction Of Changsha City Based On Random Forest

In recent years,air quality problems have occurred frequently in various parts of my country,and analysis of air quality-related indicators is of great benefit in preventing and solving air pollution problems.This paper studies the change trend of AQI index and AQI related indicators in Changsha City,Hunan Province,analyzes and predicts the concentration of major pollutants in Changsha City.The air quality in four years is explained in detail through descriptive statistics,and the annual analysis of Changsha 's AQI data is conducted to understand the air quality trends of Changsha from 2015 to 2018.Further build a spider web model of each pollutant to conduct a detailed analysis of the changing trends of various air quality indicators in Changsha.Subsequently,we use contribution analysis to draw a Pareto chart,has a preliminary understanding of the main pollutants in Changsha City within four years,and further finds out the importance of the impact of air pollutants on air quality in Changsha City through four years through random forest algorithm The IV value of each pollutant is calculated by the IV value algorithm to judge the main pollutants in Changsha from 2015 to 2018.Then,we uses cluster analysis and analysis of variance to analyze the relevant excess rate of air pollutants in Changsha in both time and space dimensions.First,through the analysis of the over-standard rate of the main pollutants in the four seasons of2018,it is obtained that the rate of over-standard in winter in Changsha City in 2018 is higher,and a systematic cluster analysis of each season in 2018 is further conducted.Then,through the analysis of over-standard rate,R cluster analysis and variance analysis of 10 air quality observation points in Changsha city,it was judged that there were significant differences in each station category.Finally,the time series model is used to predict and test the main pollutants in Changsha.The unit root test and white noise test determine that the data series of main pollutants in Changsha are converted into stationary non-white noise sequences after the second-order difference;HQIC and BIC criteria are used to model the model;finally,the model fitting effect is tested according to the DW test and the QQ chart,and a time series model is established,and the feasibility of the model is judged based on the comparison between the predicted result and the real data.Finally,the article is summarized accordingly,and a brief description of the future work is required for related issues.By analyzing the air quality of Changsha City,this paper has obtained the conclusion that the air quality status of Changsha City is better year by year,and the main pollutant in Changsha City is PM2.5,and the main polluted area is Mapoling in Changsha City.These conclusions are Changsha City.The Environmental Protection Agency has provided a strong basis for formulatingcorresponding air quality improvement strategies.