Research map / Learning theory and algorithms / Kernels and regression theory
Prediction Under Latent Factor Regression: Adaptive PCR, Interpolating Predictors and Beyond
Xin Bing, Florentina Bunea, Seth Strimas-Mackey, Marten Wegkamp
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 |
|---|---|---|---|
| increase number of features p (n=300, K=5) | PCR-K, Lasso, Ridge, GLS · Simulated factor regression | prediction risk | Improved |
| use PCR with the true K instead of Lasso when p is large | Lasso · Simulated factor regression | prediction risk | Improved |
| increase number of latent factors K | all predictors · Simulated factor regression | prediction risk | Hurt |
| add model selection (MS) by data splitting over candidate predictors | PCR, GLS, Ridge, Lasso · Simulated factor regression | prediction risk | Improved |
| choose number of components with proposed criterion (PCR-es) instead of ratio criterion (PCR-ratio) | PCR · Simulated factor regression, varying SNR | prediction risk: PCR-es estimates K consistently when SNR is larger than 2; PCR-ratio fails until SNR is greater than 4 | Improved |
| lower the signal-to-noise ratio | all predictors · Simulated factor regression, n=300, p=500, K=5 | prediction risk: Ridge best when SNR less than 2; PCR-K and PCR-es better when SNR exceeds 2 | Hurt |
| use Essential Regression predictor ER-A instead of PCR as K grows | PCR-K · Simulated Essential Regression, n=300, p=500, m=5 | prediction risk | Improved |
Closest papers by meaning
- Interpolating Predictors in High-Dimensional Factor Regression (JMLR 2022 · similarity 0.91)
- On Robustness of Principal Component Regression (NeurIPS 2019 · similarity 0.90)
- No Double Descent in Principal Component Regression: A High-Dimensional Analysis (ICML 2024 · similarity 0.90)
- On Model Identification and Out-of-Sample Prediction of PCR with Applications to Synthetic Controls (JMLR 2025 · similarity 0.89)
- Faster Principal Component Regression and Stable Matrix Chebyshev Approximation (ICML 2017 · similarity 0.87)
- Adaptive Principal Component Regression with Applications to Panel Data (NeurIPS 2023 · similarity 0.87)
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.
