A novel belief rule base representation, generation and its inference methodology

被引:110
作者
Liu, Jun [1 ,3 ]
Martinez, Luis [2 ]
Calzada, Alberto [1 ]
Wang, Hui [1 ]
机构
[1] Univ Ulster, Sch Comp & Math, Coleraine BT52 1SA, Londonderry, North Ireland
[2] Univ Jaen, Dept Comp Sci, Jaen, Spain
[3] Southwest Jiaotong Univ, Sch Math, Chengdu, Sichuan, Peoples R China
基金
中国国家自然科学基金;
关键词
Rule-based system; Belief distribution; Uncertainty; Decision-making; Inference mechanisms; Evidential reasoning approach; EVIDENTIAL REASONING APPROACH; DECISION-ANALYSIS; EXPERT-SYSTEM; ALGORITHM; MODELS;
D O I
10.1016/j.knosys.2013.08.019
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Advancement and application of rule-based systems have always been a key research area in computer-aided support for human decision making due to the fact that rule base is one of the most common frameworks for expressing various types of human knowledge in an intelligent system. In this paper, a novel rule-based representation scheme with a belief structure is proposed firstly along with its inference methodology. Such a rule base is designed with belief degrees embedded in the consequent terms as well as in the all antecedent terms of each rule, which is shown to be capable of capturing vagueness, incompleteness, uncertainty, and nonlinear causal relationships in an integrated way. The overall representation and inference framework offers a further improvement and great extension of the recently developed belief Rule base Inference Methodology (refer to as RIMER), although they still share a common scheme at the final step of inference, i.e., the evidential reasoning (ER) approach is applied to the rule combination. It is worth noting that this new extended belief rule base representation is a great extension of traditional rule base as well as fuzzy rule base by encompassing the uncertainty description in the rule antecedent and consequent. Subsequently, a simple but efficient and powerful method for automatically generating such extended belief rule base from numerical data is proposed involving neither time-consuming iterative learning procedure nor complicated rule generation mechanisms but keeping the relatively good performance, which thanks to the new features of the extended rule base with belief structures. Then some case studies in oil pipeline leak detection and software defect detection are provided to illustrate the proposed new rule base representation, generation, and inference procedure as well as demonstrate its high performance and efficiency by comparing with some existing approaches. Crown Copyright (C) 2013 Published by Elsevier B.V. All rights reserved.
引用
收藏
页码:129 / 141
页数:13
相关论文
共 37 条
[1]  
Angelov P., 2010, P IEEE INT C INT SYS, P349
[2]  
Augusto JC, 2008, INT J COMPUT INT SYS, V1, P361
[3]   Assessing the applicability of fault-proneness models across object-oriented software projects [J].
Briand, LC ;
Melo, WL ;
Wüst, J .
IEEE TRANSACTIONS ON SOFTWARE ENGINEERING, 2002, 28 (07) :706-720
[4]  
Calzada A, 2011, IEEE INT CONF FUZZY, P2638
[5]   Empirical assessment of machine learning based software defect prediction techniques [J].
Challagulla, Venkata Udaya B. ;
Bastani, Farokh B. ;
Yen, I-Ling ;
Paul, Raymond A. .
INTERNATIONAL JOURNAL ON ARTIFICIAL INTELLIGENCE TOOLS, 2008, 17 (02) :389-400
[6]   Inference analysis and adaptive training for belief rule based systems [J].
Chen, Yu-Wang ;
Yang, Jian-Bo ;
Xu, Dong-Ling ;
Zhou, Zhi-Jie ;
Tanga, Da-Wei .
EXPERT SYSTEMS WITH APPLICATIONS, 2011, 38 (10) :12845-12860
[7]  
Demsar J, 2006, J MACH LEARN RES, V7, P1
[8]   MULTILAYER FEEDFORWARD NETWORKS ARE UNIVERSAL APPROXIMATORS [J].
HORNIK, K ;
STINCHCOMBE, M ;
WHITE, H .
NEURAL NETWORKS, 1989, 2 (05) :359-366
[9]   On generating FC3 fuzzy rule systems from data using evolution strategies [J].
Jin, YC ;
von Seelen, W ;
Sendhoff, B .
IEEE TRANSACTIONS ON SYSTEMS MAN AND CYBERNETICS PART B-CYBERNETICS, 1999, 29 (06) :829-845
[10]   A belief rule-based decision support system for clinical risk assessment of cardiac chest pain [J].
Kong, Guilan ;
Xu, Dong-Ling ;
Body, Richard ;
Yang, Jian-Bo ;
Mackway-Jones, Kevin ;
Carley, Simon .
EUROPEAN JOURNAL OF OPERATIONAL RESEARCH, 2012, 219 (03) :564-573