Semi-Supervised Learning of Mixture Models

Fabio Gagliardi Cozman; Ira Cohen; Marcelo Cesar Cirelo

Semi-Supervised Learning of Mixture Models

Fabio Gagliardi Cozman, Ira Cohen, Marcelo Cesar Cirelo

Research output: Contribution to conference › Conference Paper

166 Scopus citations

Abstract

This paper analyzes the performance of semi-supervised learning of mixture models. We show that unlabeled data can lead to an increase in classification error even in situations where additional labeled data would decrease classification error. We present a mathematical analysis of this "degradation" phenomenon and show that it is due to the fact that bias may be adversely affected by unlabeled data. We discuss the impact of these theoretical results to practical situations.

Original language	American English
Pages	99-106
Number of pages	8
State	Published - 1 Dec 2003
Externally published	Yes
Event	Proceedings, Twentieth International Conference on Machine Learning - Duration: 1 Dec 2003 → …

Conference

Conference	Proceedings, Twentieth International Conference on Machine Learning
Period	1/12/03 → …

Cite this

@conference{0608b15eafe543b7bfdf939c0db6003a,

title = "Semi-Supervised Learning of Mixture Models",

abstract = "This paper analyzes the performance of semi-supervised learning of mixture models. We show that unlabeled data can lead to an increase in classification error even in situations where additional labeled data would decrease classification error. We present a mathematical analysis of this {"}degradation{"} phenomenon and show that it is due to the fact that bias may be adversely affected by unlabeled data. We discuss the impact of these theoretical results to practical situations.",

author = "Cozman, {Fabio Gagliardi} and Ira Cohen and Cirelo, {Marcelo Cesar}",

year = "2003",

month = dec,

day = "1",

language = "American English",

pages = "99--106",

note = "Proceedings, Twentieth International Conference on Machine Learning ; Conference date: 01-12-2003",

}

TY - CONF

T1 - Semi-Supervised Learning of Mixture Models

AU - Cozman, Fabio Gagliardi

AU - Cohen, Ira

AU - Cirelo, Marcelo Cesar

PY - 2003/12/1

Y1 - 2003/12/1

N2 - This paper analyzes the performance of semi-supervised learning of mixture models. We show that unlabeled data can lead to an increase in classification error even in situations where additional labeled data would decrease classification error. We present a mathematical analysis of this "degradation" phenomenon and show that it is due to the fact that bias may be adversely affected by unlabeled data. We discuss the impact of these theoretical results to practical situations.

AB - This paper analyzes the performance of semi-supervised learning of mixture models. We show that unlabeled data can lead to an increase in classification error even in situations where additional labeled data would decrease classification error. We present a mathematical analysis of this "degradation" phenomenon and show that it is due to the fact that bias may be adversely affected by unlabeled data. We discuss the impact of these theoretical results to practical situations.

UR - https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=1942452771&origin=inward

UR - https://www.scopus.com/inward/citedby.uri?partnerID=HzOxMe3b&scp=1942452771&origin=inward

M3 - Conference Paper

SP - 99

EP - 106

T2 - Proceedings, Twentieth International Conference on Machine Learning

Y2 - 1 December 2003

ER -

Semi-Supervised Learning of Mixture Models

Abstract

Conference

Other files and links

Fingerprint

Cite this