Optimal subsampling for regression with mixed‐type predictors
Open Access
- 29 July 2026
- journal article
- Published by Wiley in Canadian Journal of Statistics
Abstract
Subsampling has emerged as an appealing strategy to mitigate the computational and storage challenges imposed by large datasets. Recent subsampling techniques have shown notable computational gains for data dominated by numerical predictors. However, real‐world datasets frequently contain both numerical and categorical predictors. Traditional treatments of categorical predictors commonly use dummy coding, which can lead to substantial dimensional expansion and added complexity during subsampling. To address these challenges, we propose a new mixed‐type optimal subsampling technique (MOST) specifically designed for both numerical and categorical predictors. By enforcing combinatorial orthogonality, MOST effectively captures the informative structure of mixed‐type predictors, enabling more efficient subdata estimators. Simulation studies and real‐data applications demonstrate the superior accuracy and computational benefits of MOST compared to existing subsampling methods.Keywords
Funding Information
- National Science Foundation (DMS‐2413741)
- National Natural Science Foundation of China (12371259)
- National Natural Science Foundation of China (12501343)
This publication has 34 references indexed in Scilit:
- Orthogonal Arrays: A ReviewWIREs Computational Statistics, 2025
- Subsampling approach for least squares fitting of semi-parametric accelerated failure time models to massive survival dataStatistics and Computing, 2024
- Robust optimal subsampling based on weighted asymmetric least squaresStatistical Papers, 2023
- Distributed quantile regression for longitudinal big dataComputational Statistics, 2023
- Optimal subsampling for least absolute relative error estimators with massive dataJournal of Complexity, 2022
- Optimal subsampling for linear quantile regression modelsCanadian Journal of Statistics, 2021
- Distributed subdata selection for big data via sampling-based approachComputational Statistics & Data Analysis, 2021
- A survey of direct methods for sparse linear systemsActa Numerica, 2016
- Leveraging for big data regressionWIREs Computational Statistics, 2014
- Orthogonal Arrays1999