作　　者： ; ;

机构地区： 广州城建职业学院

出　　处： 《上海工程技术大学学报》 2018年第2期174-178,共5页

摘　　要： 研究PM_(2.5)的影响因素,基于分类回归树(CART)模型挖掘指标间的交互项和重要性,构建合理的指标体系.应用主成分分析消除解释变量间的共线性问题,建立回归模型,通过引入广义变异系数检验模型的精度.研究表明,SO_2、NO_2、CO等污染物及其交互作用是影响PM_(2.5)的重要因素,气象指标与PM_(2.5)之间有正相关关系,经济等指标对PM_(2.5)的影响有待进一步研讨. The influence factors of PM 2.5 was studied,the interaction items and importance between indexes was explored and the reasonable index system was built based on classification and regression tree （CART） model.Principal component analysis was used to eliminate the collinearity of explanatory variables and the regression model was established,and the accuracy of the model was verified by introducing generalized coefficient of variation.Research shows that pollutants such as SO 2,NO 2 and CO and their interactions are important factors affecting PM 2.5 .There is a positive correlation between meteorological indicators and PM 2.5 ,and the impact of economic indicators on PM 2.5 needs to be further discussed.

关 键 词： 回归树 交互作用 主成分回归

领　　域： []