Measure-based classifier performance evaluation

Arne Andersson; Paul Davidsson; Johan Lindén

doi:10.1016/S0167-8655(99)00084-7

Measure-based classifier performance evaluation

Arne Andersson, Paul Davidsson, Johan Lindén

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

Abstract

The concept of measure functions for classifier performance is suggested. This concept provides an alternative way of selecting and evaluating learned classifiers, and it allows us to define the learning problem as a computational problem.

Original language	English
Pages (from-to)	1165-1173
Journal	Pattern Recognition Letters
Volume	20
Issue number	11-13
DOIs	https://doi.org/10.1016/S0167-8655(99)00084-7
Publication status	Published - 1999

Subject classification (UKÄ)

Computer Science

Free keywords

Classifier performance evaluation
Generalization
Cross-validation

Access to Document

10.1016/S0167-8655(99)00084-7

Cite this

@article{b0c081995538460eb3121963744dffe4,

title = "Measure-based classifier performance evaluation",

abstract = "The concept of measure functions for classifier performance is suggested. This concept provides an alternative way of selecting and evaluating learned classifiers, and it allows us to define the learning problem as a computational problem.",

keywords = "Classifier performance evaluation, Generalization, Cross-validation",

author = "Arne Andersson and Paul Davidsson and Johan Lind{\'e}n",

year = "1999",

doi = "10.1016/S0167-8655(99)00084-7",

language = "English",

volume = "20",

pages = "1165--1173",

journal = "Pattern Recognition Letters",

issn = "0167-8655",

publisher = "Elsevier",

number = "11-13",

}

TY - JOUR

T1 - Measure-based classifier performance evaluation

AU - Andersson, Arne

AU - Davidsson, Paul

AU - Lindén, Johan

PY - 1999

Y1 - 1999

N2 - The concept of measure functions for classifier performance is suggested. This concept provides an alternative way of selecting and evaluating learned classifiers, and it allows us to define the learning problem as a computational problem.

AB - The concept of measure functions for classifier performance is suggested. This concept provides an alternative way of selecting and evaluating learned classifiers, and it allows us to define the learning problem as a computational problem.

KW - Classifier performance evaluation

KW - Generalization

KW - Cross-validation

U2 - 10.1016/S0167-8655(99)00084-7

DO - 10.1016/S0167-8655(99)00084-7

M3 - Article

SN - 0167-8655

VL - 20

SP - 1165

EP - 1173

JO - Pattern Recognition Letters

JF - Pattern Recognition Letters

IS - 11-13

ER -

Measure-based classifier performance evaluation

Abstract

Subject classification (UKÄ)

Free keywords

Access to Document

Fingerprint

Cite this