Research map / Learning theory and algorithms / Kernels and regression theory
Distributed Nonparametric Regression Imputation for Missing Response Problems with Large-scale Data
Ruoyu Wang, Miaomiao Su, Qihua Wang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use SDI (sieve) instead of KDI (kernel) when many machines and d=15 | distributed regression imputation · Simulation, d=15 | bias | Improved |
| add distributed averaging (KDI) over single-machine kernel estimator | kernel regression imputation · Simulation, d=5 | standard error | Improved |
| increase number of machines L to 500 for KDI (ablation) | KDI · Simulation, d=5 | bias | Hurt |
| increase covariate dimension from 5 to 15 for KDI (ablation) | KDI · Simulation, linear DGP | bias | Hurt |
| increase covariate dimension from 5 to 15 for SDI (ablation) | SDI · Simulation, linear DGP | bias and SE | No clear effect |
| use polynomial basis in linear DGP (zero approximation error) vs nonlinear DGP | SDI · Simulation | bias | Improved |
| make 1/pi(x) a polynomial so basis approximates it well (ablation) | SDI · Simulation, d=5 or L=10,20,50 | bias | Improved |
| raise missing rate from 25.0% to 55.9% (ablation) | SDI · Simulation, alternative missingness | standard error: 25.0% -> 55.9% | Hurt |
| add distributed computation with SDI at N=2e6 | sieve regression imputation · Simulation computing time | CPU time (s): 4003.38 -> 26.34 | Improved |
| add distributed computation with KDI on MovieLens | kernel regression imputation · MovieLens ml-25m | CPU time (s): 20896.18 -> 0.06 | Improved |
| use DWCV-selected constant c for KDI instead of fixed c=0.1 (ablation) | KDI · MovieLens ml-25m | absolute difference from benchmark: 2.072 -> 0.038 | Improved |
| increase machines to many for KDI on real data (ablation) | KDI · MovieLens ml-25m | accuracy loss | Hurt |
Closest papers by meaning
- Distribution Shift in Missing Data Imputation: A Risk-Based Perspective and Importance-Weighted Correction under MAR (NeurIPS 2026 · similarity 0.84)
- Identification and Estimation for Nonignorable Missing Data: A Data Fusion Approach (ICML 2024 · similarity 0.82)
- Diffusion Transformers for Imputation: Statistical Efficiency and Uncertainty Quantification (NeurIPS 2025 · similarity 0.82)
- Random features models: a way to study the success of naive imputation (ICML 2024 · similarity 0.82)
- Distributed High-Dimensional Quantile Regression: Estimation Efficiency and Support Recovery (ICML 2024 · similarity 0.81)
- Imputation for prediction: beware of diminishing returns. (ICLR 2025 · similarity 0.80)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
