Urban Air Quality Prediction Model Based On Improved Random Forest Algorithm

With the progress of the times and the continuous development of society,China's urban air quality pollution problem is becoming increasingly obvious,which directly affects people's health,and even the production and development of the city.Accurate and effective prediction of urban air quality is not only helpful to provide auxiliary decision-making information for urban residents,and then make scientific travel planning,but also helps to provide theoretical basis for environmental protection policies for environmental protection departments and government departments,so as to better maintain the social production and living environment.Taking urban air quality as the research object,this paper summarizes the relevant achievements of domestic and foreign scholars on urban air quality prediction,analyzes the application and improvement direction of random forest algorithm in prediction,and expounds the relevant theoretical basis of random forest algorithm;then,based on the previous research experience,it preliminarily selects urban air quality prediction factors,and combines with statistics Then,the random forest algorithm is improved from the perspective of feature selection and weighted random forest,and the parameters of the random forest algorithm are optimized to build the urban air quality prediction model,and the decision coefficient is used as the accuracy of the model The average absolute error and root mean square error are used as the evaluation indexes of the model error.Finally,the empirical analysis is carried out by using the relevant data of Chongqing from January 1,2015 to December 31,2020 every three hours,and the urban air quality indexes after 3 hours,6 hours,9 hours and 12 hours are predicted respectively,and the improved random forest algorithm and the unimproved random forest algorithm are compared In order to objectively judge the effectiveness of the urban air quality prediction model based on the improved random forest algorithm.The results show that:(1)The prediction accuracy of urban air quality prediction model based on improved random forest algorithm is higher.When predicting the urban air quality index after 3 hours,6 hours,9 hours and 12 hours,the decision coefficients of the model are 0.83,0.75,0.68 and 0.61 respectively.(2)the prediction accuracy and prediction error of the improved random forest algorithm are better than those of the unimproved random forest algorithm.In the prediction of urban air quality index after 3hours,6 hours,9 hours and 12 hours,the improved random forest algorithm compared with the unimproved random forest algorithm,the average decision coefficient increased by 4.93%,the average absolute error decreased by 5.75%,and the average root mean square error decreased by 5.36%.Finally,according to the above conclusions,the urban air quality prediction model based on the improved random forest algorithm has higher accuracy,and has better performance than the unimproved random forest algorithm,which can be used for urban air quality prediction.