基于区间分位数回归的中国空气质量指数影响因素分析

Analysis of Influencing Factors of Air Quality Index in China Based on Interval Quantile Regression

摘要: 为探究六种污染物浓度对中国空气质量指数的影响，本文采用区间分位数回归模型，对168个城市2026年6月1日至2026年6月30日的月区间数据进行实证分析，结果得到，从空气质量指数下限数据来看，$P{M}_{10}^l$ 、$S{O}_2^l$ 、$O_3^l$ 为核心影响因素；从空气质量指数上限数据来看，$P{M}_{2.5}^u$ 、$P{M}_{10}^u$ 、$O_3^u$ 为核心影响因素，且其作用强度随空气质量变化存在显著差异。预测结果表明，本文提出的基于最小最大法的区间分位数回归方法预测性能较优。

Abstract: To explore the impacts of the concentrations of six pollutants on Chinese Air Quality Index (AQI), this paper adopts an interval quantile regression model to conduct an empirical analysis on monthly interval data of 168 cities from June 1 to June 30, 2026. The results show that, in terms of the lower bound data of the Air Quality Index, $P{M}_{10}^l$ , $S{O}_2^l$ , $O_3^l$ serve as the core influencing factors; while for the upper bound data of the Air Quality Index, $P{M}_{2.5}^u$ , $P{M}_{10}^u$ , $O_3^u$ are the core influencing factors, and their effect intensities differ significantly with changes in air quality. The prediction results indicate that the proposed interval quantile regression method based on the Min-Max method achieves better prediction performance.