Research map / Learning theory and algorithms / Kernels and regression theory
Functional Linear Regression with Mixed Predictors
Daren Wang, Zifeng Zhao, Yi Yu, Rebecca Willett
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 RKHS estimator instead of best of FDA/PFFR | function-on-function regression · Scenario B, n=40,T=200,q=50, kappa=2 | R_avg (% improvement): R_avg 503.19%, R_w 100% | Improved |
| use RKHS estimator instead of FDA/PFFR | function-on-function regression · Scenario A, n=40,T=200,q=50, kappa=2 | R_avg (% improvement): R_avg 0.71, R_w 49 | No clear effect |
| use RKHS estimator instead of FDA/PFFR | function-on-function regression · Scenario B, n=20,T=100,q=20, kappa=0.5 | R_avg (% improvement): R_avg -2.41, R_w 47 | Hurt |
| use RKHS estimator instead of best baseline | function-on-function regression · Scenario B, n=20,T=100,q=20, kappa=2 | R_avg (% improvement): R_avg 36.58, R_w 100 | Improved |
| use RKHS instead of PFFR with mixed predictors | functional regression with mixed predictors · Scenario A, n=20,T=100,q=20, kappa=0.5 | R_avg (% improvement): R_avg 95.11, R_w 100 | Improved |
| use RKHS instead of PFFR with mixed predictors | functional regression with mixed predictors · Scenario B, n=40,T=200,q=50, kappa=2 | R_avg (% improvement): R_avg 827.09, R_w 100 | Improved |
| use RKHS instead of PFFR with mixed predictors | functional regression with mixed predictors · Scenario B, n=5,T=50,q=5, kappa=2 | R_avg (% improvement): R_avg 7.29, R_w 75 | Improved |
| increase SNR via kappa from 0.5 to 2 | RKHS · Function-on-function Scenario A, n=5,T=50,q=5 | nRMISE_avg x100: 15.98 -> 4.99 | Improved |
| increase q (model complexity) from 5 to 50 | RKHS · Function-on-function Scenario B, kappa=1 | nRMISE_avg x100: 10.42 -> 20.75 | Hurt |
| add vector covariates to function-on-function regression (mixed predictors) | RKHS · Kickstarter campaign forecasting, RMSE/MAE | RMSE, MAE: 19 of 28 models select no vector covariate | No clear effect |
| increase number of noise scalar predictors p to 200 (group Lasso penalty) (ablation) | RKHS · Scenario with n=20,T=100,q=20, kappa=1 | RMISE: MISE roughly linear in log(p); RKHS at p=200 beats PFFR at p=10 | Mixed |
| increase basis dimension Nb to 50 (FDA) and 30 (PFFR) (ablation) | FDA/PFFR · Scenario B q=50 | nRMISE: Nb=20 -> Nb=50 for FDA, Nb=30 for PFFR | Improved |
Closest papers by meaning
- Nonlinear function-on-function regression by RKHS (JMLR 2026 · similarity 0.90)
- Optimal Functional Bilinear Regression with Two-dimensional Functional Covariates via Reproducing Kernel Hilbert Space (JMLR 2025 · similarity 0.89)
- Statistical Optimality of Divide and Conquer Kernel-based Functional Linear Regression (JMLR 2024 · similarity 0.89)
- Transfer Learning Meets Functional Linear Regression: No Negative Transfer Under Posterior Drift (AAAI 2025 · similarity 0.87)
- On Hypothesis Transfer Learning of Functional Linear Models (ICML 2024 · similarity 0.86)
- Regression Learning with Limited Observations of Multivariate Outcomes and Features (ICML 2024 · 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.
