A Wasserstein index of dependence for random measures

09/14/2021
by   Marta Catalano, et al.
0

Nonparametric latent structure models provide flexible inference on distinct, yet related, groups of observations. Each component of a vector of d ≥ 2 random measures models the distribution of a group of exchangeable observations, while their dependence structure regulates the borrowing of information across different groups. Recent work has quantified the dependence between random measures in terms of Wasserstein distance from the maximally dependent scenario when d=2. By solving an intriguing max-min problem we are now able to define a Wasserstein index of dependence I_𝒲 with the following properties: (i) it simultaneously quantifies the dependence of d ≥ 2 random measures; (ii) it takes values in [0,1]; (iii) it attains the extreme values {0,1} under independence and complete dependence, respectively; (iv) since it is defined in terms of the underlying Lévy measures, it is possible to evaluate it numerically in many Bayesian nonparametric models for partially exchangeable data.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
11/05/2021

Why the 1-Wasserstein distance is the area between the two marginal CDFs

We elucidate why the 1-Wasserstein distance W_1 coincides with the area ...
research
02/17/2018

Domination of Sample Maxima and Related Extremal Dependence Measures

For a given d-dimensional distribution function (df) H we introduce the ...
research
11/24/2020

An AUK-based index for measuring and testing the joint dependence of a random vector

We present an index of dependence that allows one to measure the joint o...
research
07/02/2015

Correlated Random Measures

We develop correlated random measures, random measures where the atom we...
research
01/15/2018

Latent nested nonparametric priors

Discrete random structures are important tools in Bayesian nonparametric...
research
12/28/2018

Hybrid Wasserstein Distance and Fast Distribution Clustering

We define a modified Wasserstein distance for distribution clustering wh...
research
09/14/2019

Relation between non-exchangeability and measures of concordance of copulas

An investigation is presented of how a comprehensive choice of five most...

Please sign up or login with your details

Forgot password? Click here to reset