BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//132.216.98.100//NONSGML kigkonsult.se iCalcreator 2.20.4//
BEGIN:VEVENT
UID:20261121T021841EST-6337suIbbm@132.216.98.100
DTSTAMP:20261121T071841Z
DESCRIPTION:Title: Neyman-Pearson classification: parametrics and sample si
 ze requirement.\n\nAbstract: The Neyman-Pearson (NP) paradigm in binary cl
 assification seeks classifiers that achieve a minimal type II error while 
 enforcing the prioritized type I error controlled under some user-specifie
 d level alpha. This paradigm serves naturally in applications such as seve
 re disease diagnosis and spam detection\, where people have clear prioriti
 es among the two error types. Recently\, Tong\, Feng and Li (2018) propose
 d a nonparametric umbrella algorithm that adapts all scoring-type classifi
 cation methods (e.g.\, logistic regression\, support vector machines\, ran
 dom forest) to respect the given type I error (i.e.\, conditional probabil
 ity of classifying a class 0 observation as class 1 under the 0-1 coding) 
 upper bound alpha with high probability\, without specific distributional 
 assumptions on the features and the responses. Universal the umbrella algo
 rithm is\, it demands an explicit minimum sample size requirement on class
  0\, which is often the more scarce class\, such as in rare disease diagno
 sis applications. In this work\, we employ the parametric linear discrimin
 ant analysis (LDA) model and propose a new parametric thresholding algorit
 hm\, which does not need the minimum sample size requirements on class 0 o
 bservations and thus is suitable for small sample applications such as rar
 e disease diagnosis. Leveraging both the existing nonparametric and the ne
 wly proposed parametric thresholding rules\, we propose four LDA-based NP 
 classifiers\, for both low- and high-dimensional settings. On the theoreti
 cal front\, we prove NP oracle inequalities for one proposed classifier\, 
 where the rate for excess type II error benefits from the explicit paramet
 ric model assumption. Furthermore\, as NP classifiers involve a sample spl
 itting step of class 0 observations\,  we construct a new adaptive sample 
 splitting scheme that can be applied universally to NP classifiers\, and t
 his adaptive strategy reduces the type II error of these classifiers. The 
 proposed NP classifiers are implemented in the R package nproc.  \n
DTSTART:20200228T210000Z
DTEND:20200228T220000Z
LOCATION:Room 1104\, Burnside Hall\, CA\, QC\, Montreal\, H3A 0B9\, 805 rue
  Sherbrooke Ouest
SUMMARY:Yang Feng\, NYU
URL:https://www.mcgill.ca/mathstat/channels/event/yang-feng-nyu-320895
END:VEVENT
END:VCALENDAR
