A joint optimization method on parameter and structure for belief-rule-based systems

Longhao Yang, Yingming Wang, J. Liu, Luis Martinez

Research output: Contribution to journalArticle

6 Citations (Scopus)

Abstract

The belief-rule-based system (BRBS) is one of the most visible and fastest growing branches of decision support systems. As the knowledge base in the BRBS, the belief-rule-base (BRB) is required to be equipped with the optimal parameters and structure, which means the optimal value and number of parameters, respectively. Several optimization methods were therefore proposed in the past decade. However, these methods presented different limitations, such as the use of the incomplete parameter optimization model, lack of structure optimization, and so on. Moreover, it is impracticable to determine the optimal parameters and structure of a BRB using the training error because of over-fitting. The present work is focused on the joint optimization on parameter and structure for the BRB. Firstly, a simple example is utilized to illustrate and analyze the generalization capability of the BRBS under different numbers of rules, which unveils the underlying information that the BRBS with a small training error may not have superior approximation performances. Furthermore, by using the Hoeffding inequality theorem in probability theory, it is a constructive proof that the generalization error could be a better choice of criterion and measurement to determine the optimal parameters and structure of a BRB. Based on the above results, a heuristic strategy to optimize the structure of the BRB is proposed, which is followed by a parameter optimization method using the differential evolution (DE) algorithm. Finally, a joint optimization method is introduced to optimize the parameters and structure of the BRB simultaneously. In order to verify the generality and effectiveness of the proposed method, two practical case studies, namely oil pipeline leak detection and bridge risk assessment, are examined to demonstrate how the proposed method can be implemented in the BRB under disjunctive and conjunctive assumptions along with their performance comparative analysis.
LanguageEnglish
Pages220-240
Number of pages21
JournalKnowledge-Based Systems
Volume142
Issue number15
Early online date2 Dec 2017
DOIs
Publication statusPublished - 15 Feb 2018

Fingerprint

Knowledge based systems
Leak detection
Decision support systems
Risk assessment
Rule-based systems
Pipelines

Keywords

  • Belief-rule-based system
  • Parameter optimization
  • Structure optimization
  • Generalization error
  • Heuristic strategy

Cite this

Yang, Longhao ; Wang, Yingming ; Liu, J. ; Martinez, Luis. / A joint optimization method on parameter and structure for belief-rule-based systems. 2018 ; Vol. 142, No. 15. pp. 220-240.
@article{8dc16b7fe4244850b4a47f57ff40b853,
title = "A joint optimization method on parameter and structure for belief-rule-based systems",
abstract = "The belief-rule-based system (BRBS) is one of the most visible and fastest growing branches of decision support systems. As the knowledge base in the BRBS, the belief-rule-base (BRB) is required to be equipped with the optimal parameters and structure, which means the optimal value and number of parameters, respectively. Several optimization methods were therefore proposed in the past decade. However, these methods presented different limitations, such as the use of the incomplete parameter optimization model, lack of structure optimization, and so on. Moreover, it is impracticable to determine the optimal parameters and structure of a BRB using the training error because of over-fitting. The present work is focused on the joint optimization on parameter and structure for the BRB. Firstly, a simple example is utilized to illustrate and analyze the generalization capability of the BRBS under different numbers of rules, which unveils the underlying information that the BRBS with a small training error may not have superior approximation performances. Furthermore, by using the Hoeffding inequality theorem in probability theory, it is a constructive proof that the generalization error could be a better choice of criterion and measurement to determine the optimal parameters and structure of a BRB. Based on the above results, a heuristic strategy to optimize the structure of the BRB is proposed, which is followed by a parameter optimization method using the differential evolution (DE) algorithm. Finally, a joint optimization method is introduced to optimize the parameters and structure of the BRB simultaneously. In order to verify the generality and effectiveness of the proposed method, two practical case studies, namely oil pipeline leak detection and bridge risk assessment, are examined to demonstrate how the proposed method can be implemented in the BRB under disjunctive and conjunctive assumptions along with their performance comparative analysis.",
keywords = "Belief-rule-based system, Parameter optimization, Structure optimization, Generalization error, Heuristic strategy",
author = "Longhao Yang and Yingming Wang and J. Liu and Luis Martinez",
year = "2018",
month = "2",
day = "15",
doi = "10.1016/j.knosys.2017.11.039",
language = "English",
volume = "142",
pages = "220--240",
number = "15",

}

A joint optimization method on parameter and structure for belief-rule-based systems. / Yang, Longhao; Wang, Yingming; Liu, J.; Martinez, Luis.

Vol. 142, No. 15, 15.02.2018, p. 220-240.

Research output: Contribution to journalArticle

TY - JOUR

T1 - A joint optimization method on parameter and structure for belief-rule-based systems

AU - Yang, Longhao

AU - Wang, Yingming

AU - Liu, J.

AU - Martinez, Luis

PY - 2018/2/15

Y1 - 2018/2/15

N2 - The belief-rule-based system (BRBS) is one of the most visible and fastest growing branches of decision support systems. As the knowledge base in the BRBS, the belief-rule-base (BRB) is required to be equipped with the optimal parameters and structure, which means the optimal value and number of parameters, respectively. Several optimization methods were therefore proposed in the past decade. However, these methods presented different limitations, such as the use of the incomplete parameter optimization model, lack of structure optimization, and so on. Moreover, it is impracticable to determine the optimal parameters and structure of a BRB using the training error because of over-fitting. The present work is focused on the joint optimization on parameter and structure for the BRB. Firstly, a simple example is utilized to illustrate and analyze the generalization capability of the BRBS under different numbers of rules, which unveils the underlying information that the BRBS with a small training error may not have superior approximation performances. Furthermore, by using the Hoeffding inequality theorem in probability theory, it is a constructive proof that the generalization error could be a better choice of criterion and measurement to determine the optimal parameters and structure of a BRB. Based on the above results, a heuristic strategy to optimize the structure of the BRB is proposed, which is followed by a parameter optimization method using the differential evolution (DE) algorithm. Finally, a joint optimization method is introduced to optimize the parameters and structure of the BRB simultaneously. In order to verify the generality and effectiveness of the proposed method, two practical case studies, namely oil pipeline leak detection and bridge risk assessment, are examined to demonstrate how the proposed method can be implemented in the BRB under disjunctive and conjunctive assumptions along with their performance comparative analysis.

AB - The belief-rule-based system (BRBS) is one of the most visible and fastest growing branches of decision support systems. As the knowledge base in the BRBS, the belief-rule-base (BRB) is required to be equipped with the optimal parameters and structure, which means the optimal value and number of parameters, respectively. Several optimization methods were therefore proposed in the past decade. However, these methods presented different limitations, such as the use of the incomplete parameter optimization model, lack of structure optimization, and so on. Moreover, it is impracticable to determine the optimal parameters and structure of a BRB using the training error because of over-fitting. The present work is focused on the joint optimization on parameter and structure for the BRB. Firstly, a simple example is utilized to illustrate and analyze the generalization capability of the BRBS under different numbers of rules, which unveils the underlying information that the BRBS with a small training error may not have superior approximation performances. Furthermore, by using the Hoeffding inequality theorem in probability theory, it is a constructive proof that the generalization error could be a better choice of criterion and measurement to determine the optimal parameters and structure of a BRB. Based on the above results, a heuristic strategy to optimize the structure of the BRB is proposed, which is followed by a parameter optimization method using the differential evolution (DE) algorithm. Finally, a joint optimization method is introduced to optimize the parameters and structure of the BRB simultaneously. In order to verify the generality and effectiveness of the proposed method, two practical case studies, namely oil pipeline leak detection and bridge risk assessment, are examined to demonstrate how the proposed method can be implemented in the BRB under disjunctive and conjunctive assumptions along with their performance comparative analysis.

KW - Belief-rule-based system

KW - Parameter optimization

KW - Structure optimization

KW - Generalization error

KW - Heuristic strategy

U2 - 10.1016/j.knosys.2017.11.039

DO - 10.1016/j.knosys.2017.11.039

M3 - Article

VL - 142

SP - 220

EP - 240

IS - 15

ER -