摘要
采用《中国统计年鉴》的数据,利用SPSS软件在主成分分析的基础上构造了多元回归模型。根据模型产生的杠杆值、剔除学生化残差、库克距离、DFFITS统计量对我国各省的SO2排放量进行了特异数据挖掘。此外还探讨了特异值产生原因。
According to the data of China Statistical Yearbook,the multiple regression model is constructed by using SPSS software which is based on the principal component analysis.Then,through the leverage values,eliminate studentized residuals,Cook’s distance and DFFITS statistics generated by the model,the specific data of SO2 emissions of each province in China is mined.In addition,the cause for the production of specific value is discussed.
出处
《四川理工学院学报(自然科学版)》
CAS
2014年第5期96-100,共5页
Journal of Sichuan University of Science & Engineering(Natural Science Edition)
基金
四川省应用基础计划项目(2012JY0033)
关键词
SO2排放量
主成分分析
多元线性回归
特异值
SO2emissions
principal component analysis
multiple linear regression
outlier