Base belief function: an efficient method of conflict management

Dempster–Shafer evidence theory is widely used in many applications such as decision making and pattern recognition. However, Dempster’s combination rule often produces results that do not reflect the actual distribution of belief when collected evidence highly conflicts each other. In this paper, a...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Journal of ambient intelligence and humanized computing Ročník 10; číslo 9; s. 3427 - 3437
Hlavní autoři: Wang, Yunjuan, Zhang, Kezhen, Deng, Yong
Médium: Journal Article
Jazyk:angličtina
Vydáno: Berlin/Heidelberg Springer Berlin Heidelberg 01.09.2019
Springer Nature B.V
Témata:
Artificial Intelligence
Computational Intelligence
Decision making
Decision theory
Engineering
Fuzzy logic
Fuzzy sets
Hypotheses
Machine learning
Methods
Original Research
Pattern recognition
Probability
Robotics and Automation
User Interfaces and Human Computer Interaction
Conflicting evidence
Base belief function
Dempster–Shafer evidence theory
Belief function
ISSN:1868-5137, 1868-5145
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:Dempster–Shafer evidence theory is widely used in many applications such as decision making and pattern recognition. However, Dempster’s combination rule often produces results that do not reflect the actual distribution of belief when collected evidence highly conflicts each other. In this paper, a base belief function is proposed to modify the classical basic probability assignment before combination in closed-world. Base belief function focuses on making combination result intuitive especially when evidences highly conflict each other. Compared to other methods, the combination result produced by proposed method is logical and consistent with real world with less computational complexity and better performance. The advantage of base belief function is that it can avoid high conflicts between evidences and is especially suitable for the situation where the evidences appear in sequence. Several numerical examples as well as experiments using real data sets from the UCI machine learning repository for classification are employed to verify the rationality of the proposed method.
Bibliografie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:1868-5137
1868-5145
DOI:10.1007/s12652-018-1099-2