Research map / Learning theory and algorithms / Kernels and regression theory
Modular Regression: Improving Linear Models by Incorporating Auxiliary Data
Ying Jin, Dominik Rothenhäusler
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 |
|---|---|---|---|
| debias the modular estimator with semiparametric mixed-bias correction instead of naive plug-in random forest | modular regression with random forest · toy example | standard deviation vs OLS: more than 40% | Improved |
| use random forest as nuisance regressor with small sample and low noise | modular regression · simulation setting 3, sigma_z=0.1, n=200 | RMSE | Hurt |
| increase sample size from 200 to 2000 for random-forest modular regression (ablation) | modular regression with random forest · simulation setting 3 | RMSE of theta_1 | Improved |
| use modular regression with linear regression nuisance instead of plain OLS | OLS · low-dimensional simulations, 4 settings | RMSE | Improved |
| use modular regression instead of Lasso under exact conditional independence | Lasso · high-dimensional setting 1 (p=100, s=10, n=500) | parameter RMSE, bias and excess risk | Improved |
| add Lasso structure-learning step when conditional independence is approximate (ablation) | modular regression · high-dimensional setting 2 | RMSE and excess risk | Improved |
| use modular estimator versus outcome regression and orthogonal regression with ML regressors | modular regression · appendix settings (a-f) | scaled RMSE | Improved |
| naive plug-in of random forest predictions without debiasing (ablation) | modular regression · toy example | bias and SD vs OLS | Hurt |
Closest papers by meaning
- A Cluster Elastic Net for Multivariate Regression (JMLR 2018 · similarity 0.88)
- Regression Learning with Limited Observations of Multivariate Outcomes and Features (ICML 2024 · similarity 0.87)
- Fused Lasso Approach in Regression Coefficients Clustering -- Learning Parameter Heterogeneity in Data Integration (JMLR 2016 · similarity 0.87)
- Functional Linear Regression with Mixed Predictors (JMLR 2022 · similarity 0.87)
- New Insights for the Multivariate Square-Root Lasso (JMLR 2022 · similarity 0.87)
- An Improved Analysis of Alternating Minimization for Structured Multi-Response Regression (NeurIPS 2018 · similarity 0.86)
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.
