Inference for Empirical Wasserstein Distances on Finite Spaces
arXiv:1610.03287
Abstract
The Wasserstein distance is an attractive tool for data analysis but statistical inference is hindered by the lack of distributional limits. To overcome this obstacle, for probability measures supported on finitely many points, we derive the asymptotic distribution of empirical Wasserstein distances as the optimal value of a linear program with random objective function. This facilitates statistical inference (e.g. confidence intervals for sample based Wasserstein distances) in large generality. Our proof is based on directional Hadamard differentiability. Failure of the classical bootstrap and alternatives are discussed. The utility of the distributional results is illustrated on two data sets.
References in corpus (4)
- The Shortlist Method for Fast Computation of the Earth Mover's Distance and Finding Optimal Solutions to Transportation Problems
- Scalable Bayes via Barycenter in Wasserstein Space
- Inference on Directionally Differentiable Functions
- A statistical analysis of a deformation model with Wasserstein barycenters : estimation procedure and goodness of fit test
Cited by in corpus (3)
- A General Probabilistic Approach for Quantitative Assessment of LES Combustion Models
- Central Limit Theorem for empirical transportation cost in general dimension
- Lyapunov exponent and Wasserstein metric as validation tools for assessing short-time dynamics and quantitative model evaluation of large-eddy simulation