Function-on-Function Geometric Quantile Regression with a Distributed Extension
This paper proposes a novel Function-on-Function (FoF) quantile regression model for characterizing the influence of functional covariates across different parts of the conditional distribution of functional responses.We design a loss function based on geometric quantiles in infinite-dimensional spaces.This ensures that our method provides well-defined conditional quantiles for functional data, while retaining desirable statistical properties similar to those in the univariate setting.Additionally, the convergence rates of estimators are explicitly derived for both given and estimated basis functions.To address the challenges posed by large-scale functional data that are often distributed across multiple computing nodes, we further extend our framework to a distributed setting.We propose a communication-efficient smooth FoF geometric quantile regression algorithm that not only achieves computational and communication efficiency but also guarantees convergence rates comparable to those of global estimators.
Authors
- Miaomiao Yu (ORCID: https://orcid.org/0000-0002-4536-2078)
- Alan T.K. Wan
- Bingchen Li
- Yong Zhou
Institutions
- City University of Hong Kong (HK)
- East China Normal University (CN)
Publication Details
- Journal
- Statistica Sinica
- Published
- 2026-11-17
- DOI
- https://doi.org/10.5705/ss.202026.0148
- Primary Topic
- Advanced Statistical Methods and Models
- Type
- article
- Field-Weighted Citation Impact
- 0.00
Funders
- Natural Science Foundation of Shanghai
- National Natural Science Foundation of China
- Ministry of Education of the People's Republic of China
- Shanghai Municipal Education Commission