On assessing binary regression models based on ungrouped data

Chunling Lu, Yuhong Yang

Research output: Contribution to journalArticle

Abstract

Assessing a binary regression model based on ungrouped data is a commonly encountered but very challenging problem. Although tests, such as Hosmer–Lemeshow test and le Cessie–van Houwelingen test, have been devised and widely used in applications, they often have low power in detecting lack of fit and not much theoretical justification has been made on when they can work well. In this article, we propose a new approach based on a cross-validation voting system to address the problem. In addition to a theoretical guarantee that the probabilities of type I and II errors both converge to zero as the sample size increases for the new method under proper conditions, our simulation results demonstrate that it performs very well.

Original languageEnglish (US)
Pages (from-to)5-12
Number of pages8
JournalBiometrics
Volume75
Issue number1
DOIs
StatePublished - Mar 1 2019

Fingerprint

Binary Regression
Politics
Sample Size
Regression Model
Model-based
Lack of Fit
Type II error
Voting Systems
Type I error
testing
Cross-validation
Justification
Converge
Zero
Demonstrate
Simulation
sampling
methodology

Keywords

  • Goodness of fit
  • Hosmer–Lemeshow test
  • Model assessment
  • Model selection diagnostics

PubMed: MeSH publication types

  • Journal Article

Cite this

On assessing binary regression models based on ungrouped data. / Lu, Chunling; Yang, Yuhong.

In: Biometrics, Vol. 75, No. 1, 01.03.2019, p. 5-12.

Research output: Contribution to journalArticle

@article{c76555175d1b4ef1acbcf339176a1e50,
title = "On assessing binary regression models based on ungrouped data",
abstract = "Assessing a binary regression model based on ungrouped data is a commonly encountered but very challenging problem. Although tests, such as Hosmer–Lemeshow test and le Cessie–van Houwelingen test, have been devised and widely used in applications, they often have low power in detecting lack of fit and not much theoretical justification has been made on when they can work well. In this article, we propose a new approach based on a cross-validation voting system to address the problem. In addition to a theoretical guarantee that the probabilities of type I and II errors both converge to zero as the sample size increases for the new method under proper conditions, our simulation results demonstrate that it performs very well.",
keywords = "Goodness of fit, Hosmer–Lemeshow test, Model assessment, Model selection diagnostics",
author = "Chunling Lu and Yuhong Yang",
year = "2019",
month = "3",
day = "1",
doi = "10.1111/biom.12969",
language = "English (US)",
volume = "75",
pages = "5--12",
journal = "Biometrics",
issn = "0006-341X",
publisher = "Wiley-Blackwell",
number = "1",

}

TY - JOUR

T1 - On assessing binary regression models based on ungrouped data

AU - Lu, Chunling

AU - Yang, Yuhong

PY - 2019/3/1

Y1 - 2019/3/1

N2 - Assessing a binary regression model based on ungrouped data is a commonly encountered but very challenging problem. Although tests, such as Hosmer–Lemeshow test and le Cessie–van Houwelingen test, have been devised and widely used in applications, they often have low power in detecting lack of fit and not much theoretical justification has been made on when they can work well. In this article, we propose a new approach based on a cross-validation voting system to address the problem. In addition to a theoretical guarantee that the probabilities of type I and II errors both converge to zero as the sample size increases for the new method under proper conditions, our simulation results demonstrate that it performs very well.

AB - Assessing a binary regression model based on ungrouped data is a commonly encountered but very challenging problem. Although tests, such as Hosmer–Lemeshow test and le Cessie–van Houwelingen test, have been devised and widely used in applications, they often have low power in detecting lack of fit and not much theoretical justification has been made on when they can work well. In this article, we propose a new approach based on a cross-validation voting system to address the problem. In addition to a theoretical guarantee that the probabilities of type I and II errors both converge to zero as the sample size increases for the new method under proper conditions, our simulation results demonstrate that it performs very well.

KW - Goodness of fit

KW - Hosmer–Lemeshow test

KW - Model assessment

KW - Model selection diagnostics

UR - http://www.scopus.com/inward/record.url?scp=85056171759&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85056171759&partnerID=8YFLogxK

U2 - 10.1111/biom.12969

DO - 10.1111/biom.12969

M3 - Article

C2 - 30229867

AN - SCOPUS:85056171759

VL - 75

SP - 5

EP - 12

JO - Biometrics

JF - Biometrics

SN - 0006-341X

IS - 1

ER -