Confidence regions for the multinomial parameter with small sample size

TL;DR

This paper introduces a new method for constructing confidence regions for multinomial parameters with small sample sizes, applicable for any number of outcomes, ensuring controlled coverage and small volume.

Contribution

It proposes a novel non-asymptotic confidence region construction for multinomial distributions with any number of outcomes, extending beyond binomial cases.

Findings

Provides a method with controlled coverage for small samples

Applicable to any number of outcomes d>1

Ensures confidence regions with small volume

Abstract

Consider the observation of n iid realizations of an experiment with d>1 possible outcomes, which corresponds to a single observation of a multinomial distribution M(n,p) where p is an unknown discrete distribution on {1,...,d}. In many applications, the construction of a confidence region for p when n is small is crucial. This concrete challenging problem has a long history. It is well known that the confidence regions built from asymptotic statistics do not have good coverage when n is small. On the other hand, most available methods providing non-asymptotic regions with controlled coverage are limited to the binomial case d=2. In the present work, we propose a new method valid for any d>1. This method provides confidence regions with controlled coverage and small volume, and consists of the inversion of the "covering collection"' associated with level-sets of the likelihood. The behavior when d/n tends to infinity remains an interesting open problem beyond the scope of this work.