Estimating Gaussian Copulas with Missing Data

TL;DR

This paper introduces an EM-based method for estimating Gaussian copulas with missing data, improving accuracy by avoiding assumptions on marginals through semiparametric modeling.

Contribution

It presents a novel EM algorithm for Gaussian copula estimation with missing data, incorporating semiparametric modeling to relax marginal assumptions.

Findings

Joint distribution estimation is significantly more accurate.

Method outperforms existing approaches.

Semiparametric modeling enhances flexibility.

Abstract

In this work we present a rigorous application of the Expectation Maximization algorithm to determine the marginal distributions and the dependence structure in a Gaussian copula model with missing data. We further show how to circumvent a priori assumptions on the marginals with semiparametric modelling. The joint distribution learned through this algorithm is considerably closer to the underlying distribution than existing methods.