Research on Air Quality Prediction and Its Influencing Factors Based on RF Method

Air quality, as a necessary condition for human survival, directly affects people's safety, health, and living standards. With the development of industry, the problem of air pollution has reached an undeniable level. Based on this situation, this article takes the air quality index as the research object, explores its influencing factors, and constructs a prediction model. This study selected six cities, namely Beijing, Shanghai, Tianjin, Chongqing, Guangzhou, and Haikou, as the research objects. The construction of the dataset collected 15 economic factor indicators, 4 meteorological factor indicators, and 6 air pollutant indicators of the city. The KNN algorithm is used to replace outliers and fill in missing values in the dataset. After further sample expansion and indicator dimensionality reduction on the cleaned data, 654 samples and 14 related indicators were obtained. On this basis, an air quality prediction model was established under the influence of multiple factors. Firstly, when using principal component analysis to reduce the dimensionality of economic indicators, this article extracted four principal component factors with a cumulative contribution rate of 91.1%, which has a high degree of explanation for economic factors. A random effects model was established based on panel data, and the analysis results showed that green economic factors and average temperature have a significant negative impact on the Air Quality Index (AQI).