Saved in:
Bibliographic Details
Main Authors: Figueiredo, Flavio, Fernandes, José Geraldo, Silva, Jackson, Assunção, Renato M.
Format: Preprint
Published: 2023
Subjects:
Online Access:https://arxiv.org/abs/2309.16391
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913366110371840
author Figueiredo, Flavio
Fernandes, José Geraldo
Silva, Jackson
Assunção, Renato M.
author_facet Figueiredo, Flavio
Fernandes, José Geraldo
Silva, Jackson
Assunção, Renato M.
contents Copulas are powerful statistical tools for capturing dependencies across data dimensions. Applying Copulas involves estimating independent marginals, a straightforward task, followed by the much more challenging task of determining a single copulating function, $C$, that links these marginals. For bivariate data, a copula takes the form of a two-increasing function $C: (u,v)\in \mathbb{I}^2 \rightarrow \mathbb{I}$, where $\mathbb{I} = [0, 1]$. This paper proposes 2-Cats, a Neural Network (NN) model that learns two-dimensional Copulas without relying on specific Copula families (e.g., Archimedean). Furthermore, via both theoretical properties of the model and a Lagrangian training approach, we show that 2-Cats meets the desiderata of Copula properties. Moreover, inspired by the literature on Physics-Informed Neural Networks and Sobolev Training, we further extend our training strategy to learn not only the output of a Copula but also its derivatives. Our proposed method exhibits superior performance compared to the state-of-the-art across various datasets while respecting (provably for most and approximately for a single other) properties of C.
format Preprint
id arxiv_https___arxiv_org_abs_2309_16391
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle 2-Cats: 2D Copula Approximating Transforms
Figueiredo, Flavio
Fernandes, José Geraldo
Silva, Jackson
Assunção, Renato M.
Machine Learning
Artificial Intelligence
Copulas are powerful statistical tools for capturing dependencies across data dimensions. Applying Copulas involves estimating independent marginals, a straightforward task, followed by the much more challenging task of determining a single copulating function, $C$, that links these marginals. For bivariate data, a copula takes the form of a two-increasing function $C: (u,v)\in \mathbb{I}^2 \rightarrow \mathbb{I}$, where $\mathbb{I} = [0, 1]$. This paper proposes 2-Cats, a Neural Network (NN) model that learns two-dimensional Copulas without relying on specific Copula families (e.g., Archimedean). Furthermore, via both theoretical properties of the model and a Lagrangian training approach, we show that 2-Cats meets the desiderata of Copula properties. Moreover, inspired by the literature on Physics-Informed Neural Networks and Sobolev Training, we further extend our training strategy to learn not only the output of a Copula but also its derivatives. Our proposed method exhibits superior performance compared to the state-of-the-art across various datasets while respecting (provably for most and approximately for a single other) properties of C.
title 2-Cats: 2D Copula Approximating Transforms
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2309.16391