High-Dimensional Robust Factor Analysis Method Based on MRCD Estimation and Its Application
Factor analysis is one of the common multivariate statistical analysis methods.Its idea is to find a few principal factors according to the correlation between variables,and use these principal factors to describe the original variables.The traditional factor analysis method is not robust and will get unreasonable results if there are outliers in the data.Although the robust factor analysis based on MCD estimation has good anti-interference,the accuracy of MCD estimation will decrease with the increase of dimension,and this method will even lose effectiveness when the dimension is larger than the sample size.In order to carry out effective factor analysis for high-dimensional data,this paper proposes a high-dimensional robust factor analysis method based on MRCD estimation.The simulation results show that,compared with the traditional factor analysis and the MCD robust factor analysis,the MRCD high-dimensional robust factor analysis can resist the influence of outliers well and reach more reasonable conclusions under the high-dimensional data.In the empirical analysis part,this paper studies 11 coastal provinces,the results show that MRCD high-dimensional robust factor model can effectively obtain the factor analysis results of high-dimensional data;The quality of economic growth in coastal provinces is not balanced,Shanghai and Guangdong are better.
NETL
NSTL
万方数据
