Learning Bayesian networks from incomplete databases using a novel evolutionary algorithm

Man Leung WONG, Yuan Yuan GUO

Research output: Journal PublicationsJournal Article (refereed)Researchpeer-review

32 Citations (Scopus)

Abstract

This paper proposes a novel method for learning Bayesian networks from incomplete databases in the presence of missing values, which combines an evolutionary algorithm with the traditional Expectation Maximization (EM) algorithm. A data completing procedure is presented for learning and evaluating the candidate networks. Moreover, a strategy is introduced to obtain better initial networks to facilitate the method. The new method can also overcome the problem of getting stuck in sub-optimal solutions which occurs in most existing learning algorithms. The experimental results on the databases generated from several benchmark networks illustrate that the new method has better performance than some state-of-the-art algorithms. We also apply the method to a data mining problem and compare the performance of the discovered Bayesian networks with the models generated by other learning algorithms. The results demonstrate that our method outperforms other algorithms.
Original languageEnglish
Pages (from-to)368-383
Number of pages16
JournalDecision Support Systems
Volume45
Issue number2
DOIs
Publication statusPublished - 1 May 2008

Fingerprint

Bayesian networks
Evolutionary algorithms
Learning
Databases
Learning algorithms
Data mining
Benchmarking
Data Mining
Bayes Theorem
Data Base
Bayesian Networks
Evolutionary
Incomplete
Data base

Keywords

  • Bayesian networks
  • Data mining
  • evolutionary algorithms
  • machine learning

Cite this

@article{beff813814834c75a1817e47cafbf27a,
title = "Learning Bayesian networks from incomplete databases using a novel evolutionary algorithm",
abstract = "This paper proposes a novel method for learning Bayesian networks from incomplete databases in the presence of missing values, which combines an evolutionary algorithm with the traditional Expectation Maximization (EM) algorithm. A data completing procedure is presented for learning and evaluating the candidate networks. Moreover, a strategy is introduced to obtain better initial networks to facilitate the method. The new method can also overcome the problem of getting stuck in sub-optimal solutions which occurs in most existing learning algorithms. The experimental results on the databases generated from several benchmark networks illustrate that the new method has better performance than some state-of-the-art algorithms. We also apply the method to a data mining problem and compare the performance of the discovered Bayesian networks with the models generated by other learning algorithms. The results demonstrate that our method outperforms other algorithms.",
keywords = "Bayesian networks, Data mining, evolutionary algorithms, machine learning",
author = "WONG, {Man Leung} and GUO, {Yuan Yuan}",
year = "2008",
month = "5",
day = "1",
doi = "10.1016/j.dss.2008.01.002",
language = "English",
volume = "45",
pages = "368--383",
journal = "Decision Support Systems",
issn = "0167-9236",
publisher = "Elsevier",
number = "2",

}

Learning Bayesian networks from incomplete databases using a novel evolutionary algorithm. / WONG, Man Leung; GUO, Yuan Yuan.

In: Decision Support Systems, Vol. 45, No. 2, 01.05.2008, p. 368-383.

Research output: Journal PublicationsJournal Article (refereed)Researchpeer-review

TY - JOUR

T1 - Learning Bayesian networks from incomplete databases using a novel evolutionary algorithm

AU - WONG, Man Leung

AU - GUO, Yuan Yuan

PY - 2008/5/1

Y1 - 2008/5/1

N2 - This paper proposes a novel method for learning Bayesian networks from incomplete databases in the presence of missing values, which combines an evolutionary algorithm with the traditional Expectation Maximization (EM) algorithm. A data completing procedure is presented for learning and evaluating the candidate networks. Moreover, a strategy is introduced to obtain better initial networks to facilitate the method. The new method can also overcome the problem of getting stuck in sub-optimal solutions which occurs in most existing learning algorithms. The experimental results on the databases generated from several benchmark networks illustrate that the new method has better performance than some state-of-the-art algorithms. We also apply the method to a data mining problem and compare the performance of the discovered Bayesian networks with the models generated by other learning algorithms. The results demonstrate that our method outperforms other algorithms.

AB - This paper proposes a novel method for learning Bayesian networks from incomplete databases in the presence of missing values, which combines an evolutionary algorithm with the traditional Expectation Maximization (EM) algorithm. A data completing procedure is presented for learning and evaluating the candidate networks. Moreover, a strategy is introduced to obtain better initial networks to facilitate the method. The new method can also overcome the problem of getting stuck in sub-optimal solutions which occurs in most existing learning algorithms. The experimental results on the databases generated from several benchmark networks illustrate that the new method has better performance than some state-of-the-art algorithms. We also apply the method to a data mining problem and compare the performance of the discovered Bayesian networks with the models generated by other learning algorithms. The results demonstrate that our method outperforms other algorithms.

KW - Bayesian networks

KW - Data mining

KW - evolutionary algorithms

KW - machine learning

UR - http://commons.ln.edu.hk/sw_master/174

U2 - 10.1016/j.dss.2008.01.002

DO - 10.1016/j.dss.2008.01.002

M3 - Journal Article (refereed)

VL - 45

SP - 368

EP - 383

JO - Decision Support Systems

JF - Decision Support Systems

SN - 0167-9236

IS - 2

ER -