Non-parametric Clustering of Multivariate Populations with Arbitrary Sizes

TL;DR

This paper introduces a non-parametric clustering method for grouping multiple populations based on their dependence structures, applicable to panel data and using a recent statistical test for automatic cluster formation.

Contribution

It develops a novel clustering procedure that groups populations by their dependence structures using copula-based differences and a recent test statistic, adaptable to paired and panel data.

Findings

Effective clustering demonstrated on financial and insurance datasets.

Method accurately identifies groups with similar dependence structures.

Applicable to arbitrary population sizes and paired data scenarios.

Abstract

We propose a clustering procedure to group K populations into subgroups with the same dependence structure. The method is adapted to paired population and can be used with panel data. It relies on the differences between orthogonal projection coefficients of the K density copulas estimated from the K populations. Each cluster is then constituted by populations having significantly similar dependence structures. A recent test statistic from Ngounou-Bakam and Pommeret (2022) is used to construct automatically such clusters. The procedure is data driven and depends on the asymptotic level of the test. We illustrate our clustering algorithm via numerical studies and through two real datasets: a panel of financial datasets and insurance dataset of losses and allocated loss adjustment expense.