Belief rule based expert system for classification problems with new rule activation and weight calculation procedures

Chang, Leilei; Zhou, Zhijie; Yuan, You Xin; Yang, Long-Hao; Zhou, Zhiguo

doi:10.1016/j.ins.2015.12.009

Cited by 95 publications

(36 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…(1) The combinatorial explosion problem, when there are too many attributes and/or referenced values for the attributes [9];…”

Section: Advantages and Challenges By Different Types Of Brbmentioning

confidence: 99%

“…Under different assumptions, it can be classified as a conjunctive BRB or disjunctive BRB [6] [8] [9] [29]. Normally, there is no ambiguity in complex system modeling on whether a conjunctive or disjunctive model should be applied.…”

Section: Introductionmentioning

confidence: 99%

“…The disjunctive BRB requires far fewer rules in comparison with a conjunctive BRB in the same belief structure (with the same reference values for the same attributes) [6] [8] [9]. Thus, the combinatorial explosion problem can be avoided.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Indirect disjunctive belief rule base modeling using limited conjunctive rules: Two possible means

Chang

Chen

Hao

et al. 2019

International Journal of Approximate Reasoning

Self Cite

View full text Add to dashboard Cite

A traditional Belief Rule Base (BRB) is constructed under the conjunctive assumption (conjunctive BRB), which requires covering the traversal combinations of the referenced values for the attributes. Consequentially, a traditional conjunctive BRB may have to face the combinatorial explosion problem when there are too many attributes and/or referenced values for the attributes. It is difficult or at least expensive to construct a complete conjunctive BRB, while it is easy to derive one or several conjunctive rules. Comparatively, a BRB under the disjunctive assumption (disjunctive BRB) requires only covering the referenced values for the attributes instead of the traversal combination of them. Thus, the combinatorial explosion problem can be avoided. However, it is difficult to directly obtain a disjunctive BRB from either historical data or experts' knowledge. To combine the advantages of both conjunctive and disjunctive BRBs, a new approach is proposed to construct a disjunctive BRB using a limited number of conjunctive rules (insufficient to construct a complete conjunctive BRB). In the new disjunctive BRB modeling approach, each disjunctive rule is derived by quantifying its correlation with one or multiple conjunctive rules. To do so, two means for belief generation are proposed, namely, equal probability and self-organizing mapping (SOM). Two cases are studied for validating the efficiency of the proposed approach. The results by the disjunctive BRB show consistency with those derived by the conjunctive BRB as well as other approaches, which validates the efficiency of the proposed approach considering that the disjunctive BRB is constructed with only a limited number of conjunctive rules.

show abstract

“…(1) The combinatorial explosion problem, when there are too many attributes and/or referenced values for the attributes [9];…”

Section: Advantages and Challenges By Different Types Of Brbmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Indirect disjunctive belief rule base modeling using limited conjunctive rules: Two possible means

Chang

Chen

Hao

et al. 2019

International Journal of Approximate Reasoning

Self Cite

View full text Add to dashboard Cite

show abstract

“…The BRB methodology employs the informative belief structure to represent various types of information and knowledge with uncertainties and shows the capability of approximating any linear and nonlinear relationships across a wide variety of application areas. Recently, the BRB also have been applied for solving classification problem in [36][37][38]. However, a problem of BRB is the high multiplicative complexity on the number of referential values of attributes in the belief rule base [28].…”

Section: Accepted Manuscript 2 / 18mentioning

confidence: 99%

Data classification using evidence reasoning rule

Zheng

Yang

et al. 2017

Knowledge-Based Systems

184

View full text Add to dashboard Cite

Abstract-In Dempster-Shafer evidence theory (DST) based classifier design, Dempster's combination (DC) rule is commonly used as a multi-attribute classifier to combine evidence collected from different attributes. The main aim of this paper is to present a classification method using a novel combination rule i.e., the evidence reasoning (ER) rule. As an improvement of the DC rule, the newly proposed ER rule defines the reliability and weight of evidence. The former indicates the ability of attribute or its evidence to provide correct assessment for classification problem, and the latter reflects the relative important of evidence in comparison with other evidence when they need to be combined. The ER rule-based classification procedure is expatiated from evidence acquisition and estimation of evidence reliability and weight to combination of evidence. It is a purely data-driven approach without making any assumptions about the relationships between attributes and class memberships, and the specific statistic distributions of attribute data. Experiential results on five popular benchmark databases taken from University of California Irvine (UCI) machine learning database show high classification accuracy that is competitive with other classical and mainstream classifiers.

show abstract

“…Chang et al [21] suggested a rule-based expert system with a classifying approach to solve classification problems. In their research, the expert system classifier had to be optimized according to domain experts and/or previous knowledge.…”

Section: Related Workmentioning

confidence: 99%

Designing a Rule Based Expert Systems for Contact Lenses Patients

Aydilek¹,

Gümüşçü²

2018

IJITCS

View full text Add to dashboard Cite

Abstract-Expert systems that bring facts and valuable experiences together and make some deductions possible. Expression of relevant knowledge and experience in these structures may be in a set of rules. Learning problems are valid with expert systems. Therefore, they cannot add new rules and information automatically by themselves. Rules are created by human experts on the way and added upon the system. Classification datasets are collections of data commonly used in machine learning that contain and classify the previously obtained experiences. In this study, rules were obtained by using Part, NNge, Prism rule classifier algorithms, and a knowledge base of expert systems was systematically created to achieve enrichment. Enrichment and rule deduction process needs careful and sensitive attention. A combined methodology and study was revealed during this sensitive process. In this context, studies were conducted on five widely used datasets. It was aimed to reduce the redundant, conflicting, subsumed and circular rules in order to create a consistent and complete knowledge base. In this way, a methodology was developed to establish more powerful and richer contents of knowledge base that have higher quality.

show abstract

Belief rule based expert system for classification problems with new rule activation and weight calculation procedures

Cited by 95 publications

References 39 publications

Indirect disjunctive belief rule base modeling using limited conjunctive rules: Two possible means

Indirect disjunctive belief rule base modeling using limited conjunctive rules: Two possible means

Data classification using evidence reasoning rule

Designing a Rule Based Expert Systems for Contact Lenses Patients

Contact Info

Product

Resources

About