Improved methods for bandwidth selection when estimating ROC curves

Peter G. Hall; Rob J. Hyndman

doi:10.1016/S0167-7152(03)00150-0

Improved methods for bandwidth selection when estimating ROC curves

Peter G. Hall, Rob J. Hyndman^*

^*Corresponding author for this work

Research output: Contribution to journal › Article › peer-review

37 Citations (Scopus)

Abstract

The receiver operating characteristic (ROC) curve is used to describe the performance of a diagnostic test which classifies observations into two groups. We introduce new methods for selecting bandwidths when computing kernel estimates of ROC curves. Our techniques allow for interaction between the distributions of each group of observations and give substantial improvement in MISE over other proposed methods, especially when the two distributions are very different.

Original language	English
Pages (from-to)	181-189
Number of pages	9
Journal	Statistics and Probability Letters
Volume	64
Issue number	2
DOIs	https://doi.org/10.1016/S0167-7152(03)00150-0
Publication status	Published - 15 Aug 2003

Access to Document

10.1016/S0167-7152(03)00150-0

Cite this

@article{02c5e21b3eb64d0d9eb552a457a8d6bb,

title = "Improved methods for bandwidth selection when estimating ROC curves",

abstract = "The receiver operating characteristic (ROC) curve is used to describe the performance of a diagnostic test which classifies observations into two groups. We introduce new methods for selecting bandwidths when computing kernel estimates of ROC curves. Our techniques allow for interaction between the distributions of each group of observations and give substantial improvement in MISE over other proposed methods, especially when the two distributions are very different.",

keywords = "Bandwidth selection, Binary classification, Kernel estimator, MISE, ROC curve",

author = "Hall, {Peter G.} and Hyndman, {Rob J.}",

year = "2003",

month = aug,

day = "15",

doi = "10.1016/S0167-7152(03)00150-0",

language = "English",

volume = "64",

pages = "181--189",

journal = "Statistics and Probability Letters",

issn = "0167-7152",

publisher = "Elsevier B.V.",

number = "2",

}

TY - JOUR

T1 - Improved methods for bandwidth selection when estimating ROC curves

AU - Hall, Peter G.

AU - Hyndman, Rob J.

PY - 2003/8/15

Y1 - 2003/8/15

N2 - The receiver operating characteristic (ROC) curve is used to describe the performance of a diagnostic test which classifies observations into two groups. We introduce new methods for selecting bandwidths when computing kernel estimates of ROC curves. Our techniques allow for interaction between the distributions of each group of observations and give substantial improvement in MISE over other proposed methods, especially when the two distributions are very different.

AB - The receiver operating characteristic (ROC) curve is used to describe the performance of a diagnostic test which classifies observations into two groups. We introduce new methods for selecting bandwidths when computing kernel estimates of ROC curves. Our techniques allow for interaction between the distributions of each group of observations and give substantial improvement in MISE over other proposed methods, especially when the two distributions are very different.

KW - Bandwidth selection

KW - Binary classification

KW - Kernel estimator

KW - MISE

KW - ROC curve

UR - http://www.scopus.com/inward/record.url?scp=0042342663&partnerID=8YFLogxK

U2 - 10.1016/S0167-7152(03)00150-0

DO - 10.1016/S0167-7152(03)00150-0

M3 - Article

SN - 0167-7152

VL - 64

SP - 181

EP - 189

JO - Statistics and Probability Letters

JF - Statistics and Probability Letters

IS - 2

ER -

Improved methods for bandwidth selection when estimating ROC curves

Abstract

Access to Document

Other files and links

Fingerprint

Cite this