2015DOAJ (DOAJ: Directory of Open Access Journals)Requires access

Calculation and statistic test of partial correlation of general correlation measures

Wenjun Zhang

Open publisher page 5 citations

Abstract

It is well known that Pearson linear correlations between more than two attributes (nodes, taxa, variables, etc) can be adjusted to partial linear correlations for eliminating indirect between-attribute interactions of other attributes not being tested. In present study I first proposed three correlation measures, revised Dice coefficient, overlap coefficient, and proportion correlation. In addition, I proposed partial correlation measures for some correlation measures, of which Jaccard correlation, revised Dice coefficient, overlap coefficient, and point correlation are for binary attributes; Spearman rank correlation and proportion correlation are for interval value attributes. The full algorithm and Matlab codes (Pearson linear correlation is included also) are given. Users can add other general correlation measures in the Matlab codes. \n

About this research paper

What this paper is about

It is well known that Pearson linear correlations between more than two attributes (nodes, taxa, variables, etc) can be adjusted to partial linear correlations for eliminating indirect between-attribute interactions of other attributes not being tested. In present study I first proposed three correlation measures, revised Dice coefficient, overlap coefficient, and proportion correlation. In addition, I proposed partial correlation measures for some correlation measures, of which Jaccard correlation, revised Dice coefficient, overlap coefficient, and point correlation are for binary attributes; Spearman rank correlation and proportion correlation are for interval value attributes. The full algorithm and Matlab codes (Pearson linear correlation is included also) are given. Users can add other general correlation measures in the Matlab codes. \n

Why it matters

OpenAlex reports 5 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

It is well known that Pearson linear correlations between more than two attributes (nodes, taxa, variables, etc) can be adjusted to partial linear correlations for eliminating indirect between-attribute interactions of other attributes not being tested. In present study I first proposed three correlation measures, revised Dice coefficient, overlap coefficient, and proportion correlation. In addition, I proposed partial correlation measures for some correlation measures, of which Jaccard correlation, revised Dice coefficient, overlap coefficient, and point correlation are for binary attributes; Spearman rank correlation and proportion correlation are for interval value attributes. The full algorithm and Matlab codes (Pearson linear correlation is included also) are given. Users can add other general correlation measures in the Matlab codes. \n

Key concepts: Correlation, Correlation coefficient, Spearman's rank correlation coefficient, Fisher transformation, Statistics, Correlation ratio, Partial correlation, Mathematics

Related papers

Back to paper searchBrowse research topicsOriginal source
Calculation and statistic test of partial correlation of general correlation measures — Research Paper | ScholarLens