Identification of I-equivalent subnetworks in Bayesian networks to incorporate experts' knowledge

Sang Min Lee, Seoung Bum Kim

Research output: Contribution to journalArticle

Abstract

Bayesian networks (BNs) have been widely used in causal analysis because they can express the statistical relationship between significant variables. To gain superior causal analysis results, numerous studies have emphasized the importance of combining a knowledge-based approach and a data-based approach. However, combining these two approaches is a difficult task because it can reduce the effectiveness of the BN structure learning. Further, the learning schemes of BNs for computational efficiency can cause an inadequate causal analysis. To address these problems, we propose a knowledge-driven BN structure calibration algorithm for rich causal semantics. We first present an algorithm that can efficiently identify the subnetworks that can be altered to satisfy the learning condition of the BNs. We then reflect experts' knowledge to reduce erroneous causalities from the learned network. Experiments on various simulation and benchmark data sets were conducted to examine the properties of the proposed method and to compare its performance with an existing method. Further, an experimental study with real data from semiconductor fabrication plants demonstrated that the proposed method provided superior performance in improving structural accuracy.

Original languageEnglish
Article numbere12346
JournalExpert Systems
DOIs
Publication statusAccepted/In press - 2018 Jan 1

Fingerprint

Bayesian networks
Bayesian Networks
Network Structure
Structure Learning
Knowledge-based
Causality
Computational efficiency
Computational Efficiency
Semiconductors
Experimental Study
Fabrication
Calibration
Express
Semantics
Knowledge
Semiconductor materials
Benchmark
Experiment
Simulation
Experiments

Keywords

  • Bayesian networks
  • expert priors
  • inductive learning
  • knowledge representation
  • structure learning

ASJC Scopus subject areas

  • Control and Systems Engineering
  • Theoretical Computer Science
  • Computational Theory and Mathematics
  • Artificial Intelligence

Cite this

Identification of I-equivalent subnetworks in Bayesian networks to incorporate experts' knowledge. / Lee, Sang Min; Kim, Seoung Bum.

In: Expert Systems, 01.01.2018.

Research output: Contribution to journalArticle

@article{6df6ea8c6f1647d189f7c1c68c7c6030,
title = "Identification of I-equivalent subnetworks in Bayesian networks to incorporate experts' knowledge",
abstract = "Bayesian networks (BNs) have been widely used in causal analysis because they can express the statistical relationship between significant variables. To gain superior causal analysis results, numerous studies have emphasized the importance of combining a knowledge-based approach and a data-based approach. However, combining these two approaches is a difficult task because it can reduce the effectiveness of the BN structure learning. Further, the learning schemes of BNs for computational efficiency can cause an inadequate causal analysis. To address these problems, we propose a knowledge-driven BN structure calibration algorithm for rich causal semantics. We first present an algorithm that can efficiently identify the subnetworks that can be altered to satisfy the learning condition of the BNs. We then reflect experts' knowledge to reduce erroneous causalities from the learned network. Experiments on various simulation and benchmark data sets were conducted to examine the properties of the proposed method and to compare its performance with an existing method. Further, an experimental study with real data from semiconductor fabrication plants demonstrated that the proposed method provided superior performance in improving structural accuracy.",
keywords = "Bayesian networks, expert priors, inductive learning, knowledge representation, structure learning",
author = "Lee, {Sang Min} and Kim, {Seoung Bum}",
year = "2018",
month = "1",
day = "1",
doi = "10.1111/exsy.12346",
language = "English",
journal = "Expert Systems",
issn = "0266-4720",
publisher = "Wiley-Blackwell",

}

TY - JOUR

T1 - Identification of I-equivalent subnetworks in Bayesian networks to incorporate experts' knowledge

AU - Lee, Sang Min

AU - Kim, Seoung Bum

PY - 2018/1/1

Y1 - 2018/1/1

N2 - Bayesian networks (BNs) have been widely used in causal analysis because they can express the statistical relationship between significant variables. To gain superior causal analysis results, numerous studies have emphasized the importance of combining a knowledge-based approach and a data-based approach. However, combining these two approaches is a difficult task because it can reduce the effectiveness of the BN structure learning. Further, the learning schemes of BNs for computational efficiency can cause an inadequate causal analysis. To address these problems, we propose a knowledge-driven BN structure calibration algorithm for rich causal semantics. We first present an algorithm that can efficiently identify the subnetworks that can be altered to satisfy the learning condition of the BNs. We then reflect experts' knowledge to reduce erroneous causalities from the learned network. Experiments on various simulation and benchmark data sets were conducted to examine the properties of the proposed method and to compare its performance with an existing method. Further, an experimental study with real data from semiconductor fabrication plants demonstrated that the proposed method provided superior performance in improving structural accuracy.

AB - Bayesian networks (BNs) have been widely used in causal analysis because they can express the statistical relationship between significant variables. To gain superior causal analysis results, numerous studies have emphasized the importance of combining a knowledge-based approach and a data-based approach. However, combining these two approaches is a difficult task because it can reduce the effectiveness of the BN structure learning. Further, the learning schemes of BNs for computational efficiency can cause an inadequate causal analysis. To address these problems, we propose a knowledge-driven BN structure calibration algorithm for rich causal semantics. We first present an algorithm that can efficiently identify the subnetworks that can be altered to satisfy the learning condition of the BNs. We then reflect experts' knowledge to reduce erroneous causalities from the learned network. Experiments on various simulation and benchmark data sets were conducted to examine the properties of the proposed method and to compare its performance with an existing method. Further, an experimental study with real data from semiconductor fabrication plants demonstrated that the proposed method provided superior performance in improving structural accuracy.

KW - Bayesian networks

KW - expert priors

KW - inductive learning

KW - knowledge representation

KW - structure learning

UR - http://www.scopus.com/inward/record.url?scp=85054914813&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85054914813&partnerID=8YFLogxK

U2 - 10.1111/exsy.12346

DO - 10.1111/exsy.12346

M3 - Article

JO - Expert Systems

JF - Expert Systems

SN - 0266-4720

M1 - e12346

ER -