Lund University Publications

LUND UNIVERSITY LIBRARIES

A measure of dependence between two compositions

and (2009)
Abstract
A composition is a vector of positive components summing to a constant.

We consider the problem of describing the correlation between two compo-

sitions. Using a bicompositional Dirichlet distribution, we calculate a joint

correlation coefficient, based on the concept of information gain, between two

compositions. Numerical values of the joint correlation coefficient are calcu-

lated for compositions of two and three components.
Please use this url to cite or link to this publication:
author
and
organization
publishing date
type
Working paper/Preprint
publication status
published
subject
keywords
Dirichlet distribution, Correlation, Composition, Binomial coefficient differentiation, Simplex
pages
12 pages
publisher
Department of Statistics, Lund university
language
English
LU publication?
yes
id
554dfb77-d3ca-4c18-aeac-558d9ae64b1b (old id 1580926)
alternative location
http://www.stat.lu.se/publikationer/wp
date added to LUP
2016-04-04 11:43:14
date last changed
2020-08-30 02:29:27
```@misc{554dfb77-d3ca-4c18-aeac-558d9ae64b1b,
abstract     = {A composition is a vector of positive components summing to a constant.<br/><br>
We consider the problem of describing the correlation between two compo-<br/><br>
sitions. Using a bicompositional Dirichlet distribution, we calculate a joint<br/><br>
correlation coefficient, based on the concept of information gain, between two<br/><br>
compositions. Numerical values of the joint correlation coefficient are calcu-<br/><br>
lated for compositions of two and three components.},
author       = {Bergman, Jakob and Holmquist, Björn},
language     = {eng},
note         = {Working Paper},
publisher    = {Department of Statistics, Lund university},
title        = {A measure of dependence between two compositions},
url          = {https://lup.lub.lu.se/search/files/5839442/1580932.pdf},
year         = {2009},
}

```