ScholarLens Back to search
2017•Austrian Journal of StatisticsOpen access

A Modification of Linfoot's Informational Correlation Coefficient

Georgy L. Shevlyakov, Nikita Vasilevskiy

Open full text 1 citations

Abstract

Performance of the Linfoot's informational correlation coefficient is experimentally studied at the bivariate normal distribution. It is satisfactory in the case of a strong correlation and on large samples. To reduce the bias of estimation, a symmetric version of this correlation measure is proposed. On small and large samples, this modified informational correlation coefficient outperforms Linfoot's correlation measure at the bivariate normal distribution in the wide range of the correlation coefficient.

Open-access reader

About this research paper

What this paper is about

Performance of the Linfoot's informational correlation coefficient is experimentally studied at the bivariate normal distribution. It is satisfactory in the case of a strong correlation and on large samples. To reduce the bias of estimation, a symmetric version of this correlation measure is proposed. On small and large samples, this modified informational correlation coefficient outperforms Linfoot's correlation measure at the bivariate normal distribution in the wide range of the correlation coefficient.

Why it matters

OpenAlex reports 1 citations for this work. Citation counts describe recorded attention and do not establish research quality.

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

Performance of the Linfoot's informational correlation coefficient is experimentally studied at the bivariate normal distribution. It is satisfactory in the case of a strong correlation and on large samples. To reduce the bias of estimation, a symmetric version of this correlation measure is proposed. On small and large samples, this modified informational correlation coefficient outperforms Linfoot's correlation measure at the bivariate normal distribution in the wide range of the correlation coefficient.

Key concepts: Fisher transformation, Correlation ratio, Correlation coefficient, Bivariate analysis, Correlation, Mathematics, Distance correlation, Measure (data warehouse)

Related papers

Related paper 1Related paper 2Related paper 3
Back to paper searchBrowse research topicsOriginal source
A Modification of Linfoot's Informational Correlation Coefficient — Research Paper | ScholarLens