Prediction Of Air Quality Index In Fuzhou City Based On ARIMA-RF Combined Model

The problem of air pollution is becoming more and more serious all over the world,which has attracted great attention.As an evaluation standard of air quality,air quality index is very important for the research and treatment of air pollution.Therefore,it is of great theoretical significance and application value to explore the influencing factors of air quality changes and accurately predict the air quality index.ARIMA-RF combined model combines the advantages of ARIMA model which can accurately reflect the change trend of dependent variables with the advantages of random forest model which selects more variables,which can better realize the prediction.This paper will use ARIMA-RF combined model to predict air quality.As the capital of Fujian Province,Fuzhou’s air quality does not rank high in the province.Based on this,this paper analyzes the influencing factors of air quality changes in Fuzhou and uses ARIMA-RF combined model to predict the air quality index.The details are as follows:(1)The change degree of various factors affecting air quality in Fuzhou in May2017 and the proportion of air quality index in Fuzhou in December 2021 were analyzed.(2)Based on the above Fuzhou air quality index data,ARIMA model is constructed to realize the short-term prediction of Fuzhou air quality index.(3)Through the importance of Pearson correlation coefficient and random forest variables,several independent variables with strong correlation with air quality index are selected and substituted into the random forest model,and the number of decision trees and mtry value of the model are adjusted to realize the prediction of air quality index in Fuzhou.(4)Combining ARIMA model with random forest model,ARIMA-RF combined model is constructed to predict the air quality index of Fuzhou.The results show that:(1)The influence of pollutant factors on AQI is generally stronger than that of meteorological factors,of which PM2.5 has the greatest impact on AQI;(2)Among the three models,ARIMA-RF combined model has the best fitting effect,followed by random forest model,and finally ARIMA model,which shows that ARIMA-RF combined model is better than the other two models in all aspects.Finally,combined with the conclusions of this paper,this paper puts forward some views and suggestions on air governance in Fuzhou in the future.