Dealing with overdispersion in multivariate count data

TL;DR

This paper reviews likelihood-based models for overdispersion in multivariate count data, introduces a new advanced model improving variability approximation, and demonstrates its superior performance through simulation studies.

Contribution

It proposes a deeper Dirichlet-Multinomial model with a new estimation method, enhancing overdispersion modeling in high-dimensional count data.

Findings

The new model better captures observed variability.

Simulation studies confirm superior performance.

Model applicable to high-dimensional data.

Abstract

The problem of overdispersion in multivariate count data is a challenging issue. Nowadays, it covers a central role mainly due to the relevance of modern technologies data, such as Next Generation Sequencing and textual data from the web or digital collections. This work presents a comprehensive analysis of the likelihood-based models for extra-variation data proposed in the scientific literature. Particular attention will be paid to the models feasible for high-dimensional data. A new approach together with its parametric-estimation procedure is proposed. It is a deeper version of the Dirichlet-Multinomial distribution and it leads to important results allowing to get a better approximation of the observed variability. A significative comparison of these models is made through two different simulation studies that both confirm that the new model considered in this work allows to achieve the best results.