应用线性回归优化模型分析空气质量问题

Analysis of Air Quality Problems by Linear Regression Optimization Model

摘要: 针对基础的WRF-CMAQ空气预报模型的结果不理想的问题，通过对不同研究地点不同时间段的数据采集，并对官网数据的预处理，研究了在污染物排放情况不变的条件下，采用逐步回归算法分析了各气象条件与空气质量(AQI)之间的关系。根据逐步回归优化方法中需要考虑的因素，按照自变量X (气象条件)对因变量Y (AQI)作用的显著性大小，由大到小逐个引入方程中，最后根据所建立的最优回归方程的相关系数，得出影响地区空气质量的主要气象因素为气压、风速、湿度和温度，且皆与空气质量是正相关的。并且针对不同地点建立学习型线性回归优化模型，去实现对一次预报数据的修正，进而提高预报结果的准确性。

Abstract: Aiming at the problem that the results of the basic WRF-CMAQ air prediction model are not ideal, the relationship between meteorological conditions and air quality (AQI) is analyzed by stepwise regression algorithm under the condition that the pollutant emission is unchanged through data collection at different research sites and different time periods, and preprocessing of official website data. According to the factors that need to be considered in the stepwise regression optimization method and the significance of the independent variable (meteorological conditions) on the de-pendent variable (AQI), the equation is introduced one by one from large to small. Finally, according to the correlation coefficient of the established optimal regression equation, it is concluded that the main meteorological factors affecting the regional air quality are air pressure, wind speed, humidity and temperature, and they are all positively related to air quality. In addition, learning linear re-gression optimization model is established for different locations to correct the prediction data and improve the accuracy of the prediction results.