Inference methods for multivariate coefficient of variation: a novel NPC-based approach

In many fields, from psychology and biology to quality control, logistics and risk management, the coefficient of variation (CV) and in particular the multivariate coefficient of variation (MCV), can be applied as measures of variability. Consequently, several parametric and nonparametric inference...

Full description

Saved in:
Bibliographic Details
Published in:Journal of statistical computation and simulation 2025-02, Vol.95 (3), p.580-608
Main Authors: Elena, Barzizza, Nicolò, Biasetton, Riccardo, Ceccato
Format: Article
Language:English
Subjects:
Coefficient of variation
Inference
Multivariate
Multivariate analysis
permutation test
Permutations
Quality control
Risk management
Citations: Items that this one cites
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In many fields, from psychology and biology to quality control, logistics and risk management, the coefficient of variation (CV) and in particular the multivariate coefficient of variation (MCV), can be applied as measures of variability. Consequently, several parametric and nonparametric inference methods for CV and MCV have been introduced in the literature. These include a recently introduced permutation-based James-type studentized statistic for making inferences about MCV. In this article, we propose a new permutation-based solution exploiting the Non Parametric Combination (NPC) methodology. This new extension presents some major advantages: it can be applied to test both the one and two-sided ordered alternatives, it can deal with high-dimensional settings, even in the case of small sample sizes, and it can be used to deal with isotonic inference problems. In this paper, we present an extensive simulation study to address both one-way Multivariate Analysis of Variance (MANOVA)-like problems and isotonic inference problems. A case study is then provided to show a real data application of the new NPC-based solution.
ISSN:0094-9655
1563-5163
DOI:10.1080/00949655.2024.2433183