Research map / Learning theory and algorithms / Sample complexity
The Correlation-assisted Missing Data Estimator
Timothy I. Cannings, Yingying Fan
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 CAM kernel estimator instead of complete-case | kernel density/regression estimator · Density model 1 | total variation error x1e3: 6.23 -> 6.07 | Improved |
| use CAM kernel estimator instead of complete-case | kernel density estimator · Density model 2 | total variation error x1e3: 40.58 -> 40.28 | Improved |
| use CAM kernel estimator instead of complete-case | kernel regression estimator · Regression model 3 | MISE x1e3: 29.84 -> 29.51 | Improved |
| use mean imputation instead of complete-case | kernel density estimator · Density model 1 | total variation error x1e3: 6.23 -> 23.77 | Hurt |
| use PMM imputation instead of complete-case | kernel regression estimator · Regression model 3 | MISE x1e3: 29.84 -> 112.22 | Hurt |
| use PMM imputation instead of complete-case | kernel regression estimator · Regression model 1 | MISE x1e3: 4.35 -> 2.65 | Improved |
| use RF imputation instead of complete-case | kernel density estimator · Density model 2 | total variation error x1e3: 40.58 -> 38.31 | Improved |
| use CAM U-statistic instead of complete-case for bmi mean | marginal mean estimator · Terneuzen birth cohort | 95% CI width: 0.35 -> 0.22 | Improved |
| use CAM U-statistic instead of complete-case for height-weight covariance | covariance estimator · Terneuzen birth cohort | 95% CI width: 1.54 -> 0.99 | Improved |
| use regression-based linear phi_m instead of simple phi_m in CAM covariance (ablation) | CAM U-statistic · Terneuzen birth cohort | 95% CI width: 1.02 -> 0.99 | Improved |
| use regression-based linear phi_m instead of simple phi_m in CAM mean (ablation) | CAM U-statistic · Terneuzen birth cohort | 95% CI width: 0.22 -> 0.22 | No clear effect |
| use CAM local constant regression instead of complete-case | kernel regression · Brandsma school data | predictive MSE: 2.28 -> 2.25 | Improved |
Closest papers by meaning
- Identification and Estimation for Nonignorable Missing Data: A Data Fusion Approach (ICML 2024 · similarity 0.86)
- Random features models: a way to study the success of naive imputation (ICML 2024 · similarity 0.84)
- Distribution Shift in Missing Data Imputation: A Risk-Based Perspective and Importance-Weighted Correction under MAR (NeurIPS 2026 · similarity 0.83)
- Score Matching with Missing Data (ICML 2025 · similarity 0.82)
- DiffPuter: Empowering Diffusion Models for Missing Data Imputation (ICLR 2025 · similarity 0.82)
- Missing data and cluster graphs: cluster-level missingness vs variable-level missingness (NeurIPS 2026 · similarity 0.82)
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.
