Article
Computer Science, Artificial Intelligence
Yuanpeng He, Fuyuan Xiao
Summary: An improved method is proposed for addressing the conflicting management issue in the Dempster combination rule, which is crucial in multisource data fusion for applications like group decision making and target recognition. The new combination method presented in this study can handle highly conflicting environments without requiring normalization, offering convenience in computation and higher accuracy in predicting potential possibilities, especially in extreme circumstances. The validity and rationality of the proposed method are confirmed through numerical examples and real benchmark data from the UCI database.
INTERNATIONAL JOURNAL OF INTELLIGENT SYSTEMS
(2021)
Article
Computer Science, Information Systems
Ramisetty Kavya, Jabez Christopher, Subhrakanta Panda
Summary: Dempster Shafer (DS) theory is used for modeling uncertainty in information and is based on basic probability assignment. This work proposes an uncertainty measure that satisfies most of the mathematical properties and a generic set of behavioral requirements. The uncertainty value depends on the length and plausibility of the belief interval.
INFORMATION SCIENCES
(2023)
Article
Computer Science, Artificial Intelligence
Junwei Li, Baolin Xie, Yong Jin, Lin Zhou
Summary: In this study, a method based on an interval number distance model and reliability is proposed to transform objective data into a basic probability assignment. By constructing an interval number model, calculating the interval number distance, and discounting the initial basic probability assignment using static and dynamic reliability, the final basic probability assignment is obtained. Experimental results show that this method outperforms other methods in terms of classification accuracy and remains effective in an incomplete information environment.
Article
Computer Science, Artificial Intelligence
Hongfeng Long, Zhenming Peng, Yong Deng
Summary: The application of geometry in the analysis and interpretation of basic probability assignment (BPA) is a unique research direction in evidence theory. By visualizing BPA, the geometric properties and characteristics of BPA can be intuitively analyzed, and the potential features of BPA can be observed directly. Therefore, the proposed vector-based BPA visualization method has important research value.
Article
Computer Science, Artificial Intelligence
Dingbin Li, Yong Deng, Kang Hao Cheong
Summary: This study discusses the application of information quality in the frameworks of probability theory and possibility theory, introducing a method for fusing multisource information and proposing an approach applicable to evidence theory. Through a numerical example, the effectiveness of the method in pattern recognition is verified.
INTERNATIONAL JOURNAL OF INTELLIGENT SYSTEMS
(2021)
Article
Computer Science, Artificial Intelligence
Zezheng Yan, Hanping Zhao, Xiaowen Mei
Summary: The study proposes an improved combination method for conflicting evidence based on the redistribution of the basic probability assignment. The method involves calculating conflict intensity and evidence unreliability, constructing a redistribution equation for the basic probability assignment, and using information entropy to modify the basic probability assignment for more accurate results.
APPLIED INTELLIGENCE
(2022)
Article
Physics, Multidisciplinary
Jingyu Liu, Yongchuan Tang
Summary: The paper proposes a conflict data fusion method based on bBPA and evidence distance for multi-agent systems, aiming to improve the accuracy of the identification process of the MAIF system.
Article
Multidisciplinary Sciences
Shuning Wang, Yongchuan Tang
Summary: Dempster-Shafer evidence theory is commonly used for reasoning uncertain information, with generating BPA functions as the first step. A new BPA generation method based on Gaussian distribution is proposed in this paper, which involves constructing the distribution, calculating function values, data fusion, and decision-making processes. The method's feasibility and effectiveness are verified in classification problems using UCI datasets.
ARABIAN JOURNAL FOR SCIENCE AND ENGINEERING
(2022)
Article
Computer Science, Artificial Intelligence
Ming Jing, Yongchuan Tang
Summary: Dempster-Shafer evidence theory is applied to process uncertain information, but traditional methods may produce counterintuitive results. A new bBPA method is proposed to handle highly conflicting data, consistent with classical probability theory. Experimental results show the superiority of the new method in dealing with highly conflicting data.
APPLIED INTELLIGENCE
(2021)
Article
Computer Science, Information Systems
Lipeng Pan, Yong Deng
Summary: This paper extends the Dempster-Shafer evidence theory to the complex domain to effectively describe and process uncertain information in multidimensional characteristic data and periodic data with phase angle changes. It introduces the complex mass function and other basic concepts to describe uncertainty and supplements the complex Dempster rule of combination. A method to generate complex mass function and apply it to target recognition is proposed, showing improved recognition rate compared to the traditional mass function approach.
INFORMATION SCIENCES
(2022)
Article
Computer Science, Information Systems
Lingge Zhou, Huizi Cui, Xiangjun Mi, Jianfeng Zhang, Bingyi Kang
Summary: The advantages of Dempster-Shafer evidence theory (DST) in data fusion are highlighted in this paper. A novel strategy for determining the weight of evidence and assessing the reliability of comparisons is proposed. The effectiveness and practicality of the method are demonstrated through numerical experiments and application examples.
INFORMATION SCIENCES
(2022)
Article
Computer Science, Artificial Intelligence
Zhan Deng, Jianyu Wang
Summary: The study introduces a novel total uncertainty measure for the DS evidence theory framework, which can accurately quantify uncertainty in evidence and is more sensitive to changes in evidence compared to existing methods.
INTERNATIONAL JOURNAL OF INTELLIGENT SYSTEMS
(2021)
Article
Computer Science, Artificial Intelligence
Yuanpeng He, Fuyuan Xiao
Summary: To address highly conflicting evidence combinations, a new base function is proposed to alleviate conflicts and assign values to propositions based on importance. Single subset propositions are considered more crucial than multiple ones to reduce uncertainties and achieve intuitive combination results. Additionally, an averaging operation is carried out twice to prevent significant deviations between modified and original masses.
APPLIED INTELLIGENCE
(2022)
Article
Computer Science, Artificial Intelligence
Hanwen Li, Rui Cai
Summary: The paper proposes a new method to measure information quality in basic probability assignment, taking into account the influence of intersection among statements on uncertainty, and illustrates the effectiveness of this method with numerical examples. Additionally, an application in target recognition is used to demonstrate the validity of the proposed form of information quality in combining conflicting evidences.
Article
Computer Science, Artificial Intelligence
Haiyi Mao, Yong Deng
Summary: This paper addresses the issue of confidence mass allocation and proposes belief interval negation based on belief interval, showing its applications in uncertainty measurement and entropy.
APPLIED INTELLIGENCE
(2022)