2013Korean Journal of Applied StatisticsOpen access

A Comparative Study of Covariance Matrix Estimators in High-Dimensional Data

Donghyuk Lee, Jae Won Lee

Open full text 0 citations

Abstract

공분산 행렬은 다변량 통계분석에서 중요한 역할을 하고 있으며 전통적인 다변량 분석의 경우 표본 공분산 행렬이 참공분산 행렬의 추정량으로 주로 사용되었다. 하지만 변수의 수가 표본의 크기보다 훨씬 큰 고차원 데이터와 같은 경우에는 표본 공분산 행렬은 비정칙행렬이 되어 기존의 다변량 기법을 사용하는 데 적절하지 않을 수가 있다. 최근 이러한 문제점을 해결하기 위해 축소추정, 경계추정, 수정 콜레스키 분해 추정 등의 새로운 공분산 행렬의 추정량들이 제안되었다. 본 논문에서는 추정량들의 성능에 영향을 미칠 수 있는 여러 현실적인 상황들을 가정하여 모의실험을 통해 참공분산 행렬의 추정량들의 성능을 비교하였다. The covariance matrix is important in multivariate statistical analysis and a sample covariance matrix is used as an estimator of the covariance matrix. High dimensional data has a larger dimension than the sample size; therefore, the sample covariance matrix may not be suitable since it is known to perform poorly and event not invertible. A number of covariance matrix estimators have been recently proposed with three different approaches of shrinkage, thresholding, and modified Cholesky decomposition. We compare the performance of these newly proposed estimators in various situations.

Open-access reader

About this research paper

What this paper is about

공분산 행렬은 다변량 통계분석에서 중요한 역할을 하고 있으며 전통적인 다변량 분석의 경우 표본 공분산 행렬이 참공분산 행렬의 추정량으로 주로 사용되었다. 하지만 변수의 수가 표본의 크기보다 훨씬 큰 고차원 데이터와 같은 경우에는 표본 공분산 행렬은 비정칙행렬이 되어 기존의 다변량 기법을 사용하는 데 적절하지 않을 수가 있다. 최근 이러한 문제점을 해결하기 위해 축소추정, 경계추정, 수정 콜레스키 분해 추정 등의 새로운 공분산 행렬의 추정량들이 제안되었다. 본 논문에서는 추정량들의 성능에 영향을 미칠 수 있는 여러 현실적인 상황들을 가정하여 모의실험을 통해 참공분산 행렬의 추정량들의 성능을 비교하였다. The covariance matrix is important in multivariate statistical analysis and a sample covariance matrix is used as an estimator of the covariance matrix. High dimensional data has a larger dimension than the sample size; therefore, the sample covariance matrix may not be suitable since it is known to perform poorly and event not invertible. A number of covariance matrix estimators have been recently proposed with three different approaches of shrinkage, thresholding, and modified Cholesky decomposition. We compare the performance of these newly proposed estimators in various situations.

Why it matters

A significance statement is not available in the OpenAlex record.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

공분산 행렬은 다변량 통계분석에서 중요한 역할을 하고 있으며 전통적인 다변량 분석의 경우 표본 공분산 행렬이 참공분산 행렬의 추정량으로 주로 사용되었다. 하지만 변수의 수가 표본의 크기보다 훨씬 큰 고차원 데이터와 같은 경우에는 표본 공분산 행렬은 비정칙행렬이 되어 기존의 다변량 기법을 사용하는 데 적절하지 않을 수가 있다. 최근 이러한 문제점을 해결하기 위해 축소추정, 경계추정, 수정 콜레스키 분해 추정 등의 새로운 공분산 행렬의 추정량들이 제안되었다. 본 논문에서는 추정량들의 성능에 영향을 미칠 수 있는 여러 현실적인 상황들을 가정하여 모의실험을 통해 참공분산 행렬의 추정량들의 성능을 비교하였다. The covariance matrix is important in multivariate statistical analysis and a sample covariance matrix is used as an estimator of the covariance matrix. High dimensional data has a larger dimension than the sample size; therefore, the sample covariance matrix may not be suitable since it is known to perform poorly and event not invertible. A number of covariance matrix estimators have been recently proposed with three different approaches of shrinkage, thresholding, and modified Cholesky decomposition. We compare the performance of these newly proposed estimators in various situations.

Key concepts: Cholesky decomposition, Estimation of covariance matrices, Covariance matrix, Scatter matrix, Covariance intersection, Mathematics, Rational quadratic covariance function, Matérn covariance function

Related papers

Back to paper searchBrowse research topicsOriginal source
A Comparative Study of Covariance Matrix Estimators in High-Dimensional Data — Research Paper | ScholarLens