Classification using Hierarchical Naive Bayes models

Helge  Langseth; Thomas  Dyhre Nielsen

doi:10.1007/s10994-006-6136-2

Classification using Hierarchical Naive Bayes models

Helge Langseth, Thomas Dyhre Nielsen

Institut for Datalogi

Publikation: Bidrag til tidsskrift › Tidsskriftartikel › Forskning › peer review

70 Citationer (Scopus)

1002 Downloads (Pure)

Abstract

Udgivelsesdato: MAY

Originalsprog	Engelsk
Tidsskrift	Machine Learning
Vol/bind	63
Udgave nummer	2
Sider (fra-til)	135-159
ISSN	0885-6125
DOI	https://doi.org/10.1007/s10994-006-6136-2
Status	Udgivet - 2006

Adgang til dokumentet

10.1007/s10994-006-6136-2

LangsethNielsen-HNBIndsendt manuskript, 326 KB

http://dx.doi.org/10.1007/s10994-006-6136-2

AUB Link

Søg efter materialet i Aalborg Universitetsbiblioteks søgemaskine

Citationsformater

@article{a99f4830a08e11db8ed6000ea68e967b,

title = "Classification using Hierarchical Naive Bayes models",

abstract = "Classification problems have a long history in the machine learning literature. One of the simplest, and yet most consistently well-performing set of classifiers is the Na{\"i}ve Bayes models. However, an inherent problem with these classifiers is the assumption that all attributes used to describe an instance are conditionally independent given the class of that instance. When this assumption is violated (which is often the case in practice) it can reduce classification accuracy due to “information double-counting” and interaction omission. In this paper we focus on a relatively new set of models, termed Hierarchical Na{\"i}ve Bayes models. Hierarchical Na{\"i}ve Bayes models extend the modeling flexibility of Na{\"i}ve Bayes models by introducing latent variables to relax some of the independence statements in these models. We propose a simple algorithm for learning Hierarchical Na{\"i}ve Bayes models in the context of classification. Experimental results show that the learned models can significantly improve classification accuracy as compared to other frameworks.",

keywords = "Classification, Na{\"i}ve Bayes models",

author = "Helge Langseth and {Dyhre Nielsen}, Thomas",

year = "2006",

doi = "10.1007/s10994-006-6136-2",

language = "English",

volume = "63",

pages = "135--159",

journal = "Machine Learning",

issn = "0885-6125",

publisher = "Springer",

number = "2",

}

TY - JOUR

T1 - Classification using Hierarchical Naive Bayes models

AU - Langseth, Helge

AU - Dyhre Nielsen, Thomas

PY - 2006

Y1 - 2006

N2 - Classification problems have a long history in the machine learning literature. One of the simplest, and yet most consistently well-performing set of classifiers is the Naïve Bayes models. However, an inherent problem with these classifiers is the assumption that all attributes used to describe an instance are conditionally independent given the class of that instance. When this assumption is violated (which is often the case in practice) it can reduce classification accuracy due to “information double-counting” and interaction omission. In this paper we focus on a relatively new set of models, termed Hierarchical Naïve Bayes models. Hierarchical Naïve Bayes models extend the modeling flexibility of Naïve Bayes models by introducing latent variables to relax some of the independence statements in these models. We propose a simple algorithm for learning Hierarchical Naïve Bayes models in the context of classification. Experimental results show that the learned models can significantly improve classification accuracy as compared to other frameworks.

AB - Classification problems have a long history in the machine learning literature. One of the simplest, and yet most consistently well-performing set of classifiers is the Naïve Bayes models. However, an inherent problem with these classifiers is the assumption that all attributes used to describe an instance are conditionally independent given the class of that instance. When this assumption is violated (which is often the case in practice) it can reduce classification accuracy due to “information double-counting” and interaction omission. In this paper we focus on a relatively new set of models, termed Hierarchical Naïve Bayes models. Hierarchical Naïve Bayes models extend the modeling flexibility of Naïve Bayes models by introducing latent variables to relax some of the independence statements in these models. We propose a simple algorithm for learning Hierarchical Naïve Bayes models in the context of classification. Experimental results show that the learned models can significantly improve classification accuracy as compared to other frameworks.

KW - Classification

KW - Naïve Bayes models

U2 - 10.1007/s10994-006-6136-2

DO - 10.1007/s10994-006-6136-2

M3 - Journal article

SN - 0885-6125

VL - 63

SP - 135

EP - 159

JO - Machine Learning

JF - Machine Learning

IS - 2

ER -

Classification using Hierarchical Naive Bayes models

Abstract

Adgang til dokumentet

AUB Link

Fingeraftryk

Citationsformater