Research map / Probabilistic and scientific modeling / Variational and generative sampling
LinCDE: Conditional Density Estimation via Lindsey's Method
Zijun Gao, Trevor Hastie
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 LinCDE boosting instead of distribution boosting | conditional density estimators · Height (age 14-17) | negative log-likelihood: 4.21 -> 3.61 | Improved |
| Use LinCDE boosting instead of QRF | conditional density estimators · Bone density | negative log-likelihood: -1.67 -> -1.89 | Improved |
| Use LinCDE boosting instead of QRF | conditional density estimators · Geyser | negative log-likelihood: 1.55 -> 1.16 | Improved |
| Use LinCDE boosting instead of QRF with more covariates | conditional density estimators · Height (age 14-40) | negative log-likelihood: 3.30 -> 3.19 | Improved |
| Use LinCDE boosting instead of quantile-based methods for quantile loss | LinCDE boosting · LGD | 50% quantile loss: 0.218 -> 0.212 | No clear effect |
| Use LinCDE boosting instead of QRF for quantile loss | LinCDE boosting · LGGMD | 50% quantile loss: 0.246 -> 0.246 | No clear effect |
| Use LinCDE boosting instead of distribution boosting for computation | LinCDE boosting · LGD | training time (s): 13 -> 4.8 | Improved |
| Use LinCDE boosting instead of QRF for training speed | LinCDE boosting · LGD | training time (s): 1.3 -> 4.8 | Hurt |
| Use LinCDE boosting instead of QRF for 90% PI width | LinCDE boosting · LGD | 90% PI width: 2.02 -> 1.84 | Improved |
| Use LinCDE boosting instead of QRF for 90% PI width | LinCDE boosting · LGGMD | 90% PI width: 1.76 -> 1.78 | No clear effect |
| Use LinCDE boosting instead of QRF for variable importance | LinCDE boosting · LGD | importance on influential covariates: 40% -> 87% | Improved |
| Add LinCDE boosting (ensemble) in place of single LinCDE trees | LinCDE · toy locally Gaussian | estimation band width: narrower by around a half | Improved |
Closest papers by meaning
- Conditional Density Estimation with Histogram Trees (NeurIPS 2024 · similarity 0.87)
- The Expxorcist: Nonparametric Graphical Models Via Conditional Exponential Densities (NeurIPS 2017 · similarity 0.85)
- Deconvolutional Density Network: Modeling Free-Form Conditional Distributions (AAAI 2022 · similarity 0.84)
- Partition Tree: Conditional Density Estimation over General Outcome Spaces (NeurIPS 2026 · similarity 0.84)
- Density Ratio Estimation with Conditional Probability Paths (ICML 2025 · similarity 0.84)
- SD-KDE: Score-Debiased Kernel Density Estimation (NeurIPS 2025 · 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.
