pipette
ENEnglish

Weighted Power Fr\'echet Means in Metric Spaces with Curvature Bounded Above

Christof Sch\"otz

Preprint

En palabras de los autores

We establish non-asymptotic risk bounds for power Fr\'echet means in geodesic metric spaces with curvature bounded above. The observations form weighted, possibly infinite sequences of independent random variables whose laws and means may differ. We treat three settings. For -Fr\'echet means in Hadamard spaces, we obtain a sharp mean squared error bound that becomes an identity in Hilbert spaces. For -Fr\'echet means in CAT() spaces with , we establish variance and Wasserstein contraction inequalities with optimal constants depending on the circumradius of the closed convex data domain, and obtain mean squared error bounds throughout the maximal range . For -Fr\'echet means in Hadamard spaces, , we derive finite risk bounds under weighted -moment conditions, allowing even an infinite -moment of the population mixture. The proofs combine variance, quadruple, and contraction inequalities with a leave-one-out stability technique. Applications give prior and posterior bounds for Dirichlet-process Fr\'echet means and finite-sample guarantees for local constant Fr\'echet regression. The regression results require no response-space entropy condition and replace density smoothness assumptions common in earlier work with transport smoothness; for , the bounds remain finite even when the responses have infinite variance.

Resultado principalEl resumen no menciona limitaciones.

Apareció: viernes, 2 de octubre. arXiv. Preprint, todavía sin revisión por pares.