An optimal transportation approach for assessing almost stochastic order

2017/05/04 by E. del Barrio, del Barrio, E., Juan A. Cuesta‐Albertos +3 · 3 citations
Decision Sciences · #FOS: Computer and information sciences #Methodology (stat.ME) #Risk and Portfolio Optimization

paper · pdf · doi:10.48550/arxiv.1705.01788

openalex publication_date 2017/05/04 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

When stochastic dominance F≤_stG does not hold, we can improve agreement to stochastic order by suitably trimming both distributions. In this work we consider the L₂-Wasserstein distance, \mathcal W₂, to stochastic order of these trimmed versions. Our characterization for that distance naturally leads to consider a \mathcal W₂-based index of disagreement with stochastic order, ε_{\mathcal W2}(F,G). We provide asymptotic results allowing to test H₀: ε_{\mathcal W2}(F,G)≥ ε₀ vs H_a: ε_{\mathcal W2}(F,G)

Cited by

• ManufactuBERT: Efficient Continual Pretraining for Manufacturing
• Almost stochastic dominance via optimal transport
• The AI Gap: How Socioeconomic Status Affects Language Technology Interactions

Related

• Almost stochastic dominance via optimal transport
• Statistical inference of convex order by Wasserstein projection
• Estimating processes in adapted Wasserstein distance
• Heterogeneous Wasserstein Discrepancy for Incomparable Distributions
• Tests for Almost Stochastic Dominance