Shift-adjusted Neyman–Pearson classifiers via single index modeling (SACSIM)
Open Access
- 1 July 2026
- journal article
- Published by Oxford University Press (OUP) in Biometrics
- Vol. 82 (3)
Abstract
Neyman–Pearson (NP) classifiers, which aim to maximize the clinical benefit while adhering to risk constraints, are crucial in many practical fields, including early cancer detection. However, applying these classifiers can be challenging due to discrepancies between the data distributions of the source and target populations. The potential impact can be disproportionately severe for under-represented groups. We propose a semi-parametric model-based approach for adapting NP classifier decision rules to different populations while equitably controlling classification errors specific to clinical applications. Our method involves a shift-adjustment strategy that leverages a small unlabeled sample from the target population, along with minimal auxiliary information and the labeled source data. This approach enhances the applicability of the learned decision rules and ensures they are consistently tailored for the target population. We demonstrate the performance through theoretical studies and simulations and illustrate the approach with an example of a prostate cancer study.Keywords
Funding Information
- National Institutes of Health (R01 CA289668)
- National Institutes of Health (R01 CA236558)
- National Institutes of Health (U24 CA288185)
- National Institutes of Health (U24 CA086368)
This publication has 24 references indexed in Scilit:
- Development and Validation of an 18-Gene Urine Test for High-Grade Prostate CancerJAMA Oncology, 2024
- Association between prediagnostic prostate‐specific antigen and prostate cancer probability in Black and non‐Hispanic White menCancer, 2023
- Simultaneous selection and incorporation of consistent external aggregate informationStatistics in Medicine, 2023
- Association Between Combined TMPRSS2:ERG and PCA3 RNA Urinary Testing and Detection of Aggressive Prostate CancerJAMA Oncology, 2017
- Prevalence and Prognostic Significance of PTEN Loss in African-American and European-American Men Undergoing Radical ProstatectomyEuropean Urology, 2017
- A unifying view on dataset shift in classificationPattern Recognition, 2012
- Direct importance estimation for covariate shift adaptationAnnals of the Institute of Statistical Mathematics, 2008
- Rates of strong uniform consistency for multivariate kernel density estimatorsAnnales de l'Institut Henri Poincaré, Probabilités et Statistiques, 2002
- A Probabilistic Theory of Pattern Recognition1996
- Semiparametric least squares (SLS) and weighted SLS estimation of single-index modelsJournal of Econometrics, 1993