Font Size: a A A

The Application Of Robust Principal Component Regression Method In Evaluating The Data Quality Of Economic Growth In China

Posted on:2021-04-27Degree:MasterType:Thesis
Country:ChinaCandidate:S R DengFull Text:PDF
GTID:2439330647957067Subject:Applied statistics
Abstract/Summary:PDF Full Text Request
The quality of China's economic growth data has always been a closely concerned issue at home and abroad.Some scholars' doubts about China's economic growth data have damaged China's international image.Therefore,it is of great practical significance to find a scientific and reasonable data quality assessment method.Therefore,the robust principal component regression method is used to evaluate the quality of China's economic data in this paper.This method can not only make the obtained regression estimation not strongly affected by outliers,but also identify outliers better.In the stage of robust principal component analysis,KSD+Rocke estimation method is used to obtain the robust mean vector and covariance matrix.In the stage of robust regression analysis,MM estimation method is used to obtain the robust regression parameter estimation.After 31 provincial level administrative units in 2018 growth rate of GDP growth and 14 related indicators of the empirical analysis,the following conclusion: although a small number of indicators of growth regression coefficient does not accord with expectations,but through the analysis of the reasons for this phenomenon can be found,so comprehensive can think that the regression coefficient of symbol is consistent with economic laws,China's economic growth data is quality assured;The diagnosis method of outlier points shows that the economic growth data of all provinces are reliable on the whole,but there are some anomalies.The economic growth data of tianjin,hebei,jilin,heilongjiang and chongqing are likely to be underestimated,and the economic growth data of xinjiang uygur autonomous region is of doubtful quality.Robust principal component regression method is superior to traditional principal component regression method in the recognition of abnormal points.
Keywords/Search Tags:Robust principal component regression, Outlier point diagnosis, Data quality
PDF Full Text Request
Related items