Research map / Learning theory and algorithms / Prediction and decision losses
Neyman-Pearson classification: parametrics and sample size requirement
Xin Tong, Lucy Xia, Jiacheng Wang, Yang Feng
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 NP-sLDA instead of classical sLDA | NP-sLDA · Ex 3 | type I violation rate: .764 -> .068 | Improved |
| use NP-sLDA instead of classical sLDA | NP-sLDA · Ex 4 | type I violation rate: 1.000 -> .073 | Improved |
| use NP-sLDA instead of classical sLDA | NP-sLDA · Ex 3 | type II error (mean): .104 -> .189 | Hurt |
| use NP-sLDA instead of NP-svm | NP-sLDA · Ex 3 | type II error (mean): .621 -> .189 | Improved |
| use NP-sLDA instead of NP-penlog | NP-sLDA · Ex 3 | type II error (mean): .205 -> .189 | Improved |
| use NP-sLDA instead of NP-penlog | NP-sLDA · Ex 5 | type II error (mean): .334 -> .332 | No clear effect |
| use NP-sLDA nonparametric threshold instead of parametric pNP-sLDA | NP-sLDA · Ex 4 | type II error (mean): .824 -> .246 | Improved |
| use parametric threshold pNP-sLDA instead of NP umbrella | pNP-sLDA · Ex 4 | type I violation rate: .073 -> .000 | Hurt |
| add adaptive sample splitting instead of half-half split (ablation) | NP classifiers · Example 6a (d=1000, N0=100, varying n1/N0) | type II error | Improved |
| increase class 1 to class 0 ratio from 1 to 16 (ablation) | NP-penlog and NP-sLDA · Example 6a | type II error | Improved |
| use pNP over NP umbrella at small N0 | pNP-LDA and pNP-sLDA · Example 1 | type II error | Improved |
| increase dimension d | NP umbrella over parametric threshold · Example 1c, 1d | type II error | Mixed |
Closest papers by meaning
- Non-splitting Neyman-Pearson Classifiers (JMLR 2024 · similarity 0.95)
- Neyman-Pearson Classification under High-Dimensional Settings (JMLR 2016 · similarity 0.93)
- Neyman-Pearson Classification under Both Null and Alternative Distributions Shift (ICLR 2026 · similarity 0.84)
- Sparse Semiparametric Discriminant Analysis for High-dimensional Zero-inflated Data (JMLR 2025 · similarity 0.83)
- Concentration and excess risk bounds for imbalanced classification with synthetic oversampling (NeurIPS 2025 · similarity 0.82)
- Non-parametric classification via expand-and-sparsify representation (NeurIPS 2024 · similarity 0.81)
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.
