Building Bayesian network based expert systems from rules

Research output: Chapter in Book/Report/Conference proceedingConference contribution

3 Citations (Scopus)

Abstract

Combining expert knowledge and user explanation with automated reasoning in domains with uncertain information poses significant challenges in terms of representation and reasoning mechanisms. In particular, reasoning structures understandable and usable by humans are often different from the ones used for automated reasoning and data mining systems. Rules with certainty factors represent one possible way to express domain knowledge and build expert system that can deal with uncertainty. Although convenient to humans, this approach has limitations in accurately modeling the domain. Alternatively, a Bayesian Network allows accurate modeling of a domain and automated reasoning but its inference is less intuitive to humans. In this paper, we propose a method to combine these two frameworks to build Bayesian Networks from rules and derive user understandable explanations in terms of these rules. Expert specified rules are augmented with importance parameters for antecedents and are used to derive probabilistic bounds for the Bayesian Network's conditional probability table. The partial structure constructed from the rules is fully learned from the data. The paper also discusses methods for using the rules to provide user understandable explanations, identify incorrect rules, suggest new rules and perform incremental learning.

Original languageEnglish
Title of host publication2011 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2011 - Conference Digest
Pages3002-3008
Number of pages7
DOIs
Publication statusPublished - 23 Dec 2011
Externally publishedYes
Event2011 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2011 - Anchorage, AK, United States
Duration: 9 Oct 201112 Oct 2011

Other

Other2011 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2011
CountryUnited States
CityAnchorage, AK
Period9/10/1112/10/11

Fingerprint

Bayesian networks
Expert systems
Data mining

Keywords

  • Bayesian Networks
  • Certainty factors
  • expert systems

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Control and Systems Engineering
  • Human-Computer Interaction

Cite this

Thirumuruganathan, S., & Huber, M. (2011). Building Bayesian network based expert systems from rules. In 2011 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2011 - Conference Digest (pp. 3002-3008). [6084157] https://doi.org/10.1109/ICSMC.2011.6084157

Building Bayesian network based expert systems from rules. / Thirumuruganathan, Saravanan; Huber, Manfred.

2011 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2011 - Conference Digest. 2011. p. 3002-3008 6084157.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Thirumuruganathan, S & Huber, M 2011, Building Bayesian network based expert systems from rules. in 2011 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2011 - Conference Digest., 6084157, pp. 3002-3008, 2011 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2011, Anchorage, AK, United States, 9/10/11. https://doi.org/10.1109/ICSMC.2011.6084157
Thirumuruganathan S, Huber M. Building Bayesian network based expert systems from rules. In 2011 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2011 - Conference Digest. 2011. p. 3002-3008. 6084157 https://doi.org/10.1109/ICSMC.2011.6084157
Thirumuruganathan, Saravanan ; Huber, Manfred. / Building Bayesian network based expert systems from rules. 2011 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2011 - Conference Digest. 2011. pp. 3002-3008
@inproceedings{ded7e53b342d4ad49b1aa0ee81ff2899,
title = "Building Bayesian network based expert systems from rules",
abstract = "Combining expert knowledge and user explanation with automated reasoning in domains with uncertain information poses significant challenges in terms of representation and reasoning mechanisms. In particular, reasoning structures understandable and usable by humans are often different from the ones used for automated reasoning and data mining systems. Rules with certainty factors represent one possible way to express domain knowledge and build expert system that can deal with uncertainty. Although convenient to humans, this approach has limitations in accurately modeling the domain. Alternatively, a Bayesian Network allows accurate modeling of a domain and automated reasoning but its inference is less intuitive to humans. In this paper, we propose a method to combine these two frameworks to build Bayesian Networks from rules and derive user understandable explanations in terms of these rules. Expert specified rules are augmented with importance parameters for antecedents and are used to derive probabilistic bounds for the Bayesian Network's conditional probability table. The partial structure constructed from the rules is fully learned from the data. The paper also discusses methods for using the rules to provide user understandable explanations, identify incorrect rules, suggest new rules and perform incremental learning.",
keywords = "Bayesian Networks, Certainty factors, expert systems",
author = "Saravanan Thirumuruganathan and Manfred Huber",
year = "2011",
month = "12",
day = "23",
doi = "10.1109/ICSMC.2011.6084157",
language = "English",
isbn = "9781457706523",
pages = "3002--3008",
booktitle = "2011 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2011 - Conference Digest",

}

TY - GEN

T1 - Building Bayesian network based expert systems from rules

AU - Thirumuruganathan, Saravanan

AU - Huber, Manfred

PY - 2011/12/23

Y1 - 2011/12/23

N2 - Combining expert knowledge and user explanation with automated reasoning in domains with uncertain information poses significant challenges in terms of representation and reasoning mechanisms. In particular, reasoning structures understandable and usable by humans are often different from the ones used for automated reasoning and data mining systems. Rules with certainty factors represent one possible way to express domain knowledge and build expert system that can deal with uncertainty. Although convenient to humans, this approach has limitations in accurately modeling the domain. Alternatively, a Bayesian Network allows accurate modeling of a domain and automated reasoning but its inference is less intuitive to humans. In this paper, we propose a method to combine these two frameworks to build Bayesian Networks from rules and derive user understandable explanations in terms of these rules. Expert specified rules are augmented with importance parameters for antecedents and are used to derive probabilistic bounds for the Bayesian Network's conditional probability table. The partial structure constructed from the rules is fully learned from the data. The paper also discusses methods for using the rules to provide user understandable explanations, identify incorrect rules, suggest new rules and perform incremental learning.

AB - Combining expert knowledge and user explanation with automated reasoning in domains with uncertain information poses significant challenges in terms of representation and reasoning mechanisms. In particular, reasoning structures understandable and usable by humans are often different from the ones used for automated reasoning and data mining systems. Rules with certainty factors represent one possible way to express domain knowledge and build expert system that can deal with uncertainty. Although convenient to humans, this approach has limitations in accurately modeling the domain. Alternatively, a Bayesian Network allows accurate modeling of a domain and automated reasoning but its inference is less intuitive to humans. In this paper, we propose a method to combine these two frameworks to build Bayesian Networks from rules and derive user understandable explanations in terms of these rules. Expert specified rules are augmented with importance parameters for antecedents and are used to derive probabilistic bounds for the Bayesian Network's conditional probability table. The partial structure constructed from the rules is fully learned from the data. The paper also discusses methods for using the rules to provide user understandable explanations, identify incorrect rules, suggest new rules and perform incremental learning.

KW - Bayesian Networks

KW - Certainty factors

KW - expert systems

UR - http://www.scopus.com/inward/record.url?scp=83755228674&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=83755228674&partnerID=8YFLogxK

U2 - 10.1109/ICSMC.2011.6084157

DO - 10.1109/ICSMC.2011.6084157

M3 - Conference contribution

AN - SCOPUS:83755228674

SN - 9781457706523

SP - 3002

EP - 3008

BT - 2011 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2011 - Conference Digest

ER -