Combination of Transferable Classification With Multisource Domain Adaptation Based on Evidential Reasoning

Liu, Zhun-ga; Huang, Linqing; Zhou, Kuang; Denœux, Thierry

doi:10.1109/tnnls.2020.2995862

Cited by 80 publications

(18 citation statements)

References 47 publications

(72 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Parameter is tuned to maximize a "benefit value" that depends on the cardinality of the sets. Liu et al use a similar approach in [27] but implement a strategy that selects either a single class, or a pair of classes.…”

Section: Related Workmentioning

confidence: 99%

Partial classification in the belief function framework

Denœux

2021

Knowledge-Based Systems

View full text Add to dashboard Cite

Partial, or set-valued classification assigns instances to sets of classes, making it possible to reduce the probability of misclassification while still providing useful information. This paper reviews approaches to partial classification based on the Dempster-Shafer theory of belief functions. To define the utility of set-valued predictions, we propose to extend the utility matrix using an Ordered Weighted Average operator, allowing us to model the decision maker's attitude towards imprecision using a single parameter. Various decision criteria are analyzed comprehensively. In particular, two main strategies are distinguished: partial classification based on complete preorders among partial assignments, and partial preorders among complete assignments. Experiments with UCI and simulated Gaussian data sets show the superiority of partial classification in terms of average utility, as compared to single-class assignment and classification with rejection.

show abstract

Section: Related Workmentioning

confidence: 99%

Partial classification in the belief function framework

Denœux

2021

Knowledge-Based Systems

View full text Add to dashboard Cite

show abstract

“…Chen, Wang, Ying, & Cao, 2021;Huang et al, 2021;Wang, Liu, Zhao, Guo, & Terzijae, 2020;Wang, Wei, et al, 2020), complex networks (T. Wen & Cheong, 2021), target recognition (Z. Wen, Liu, Zhang, & Pan, 2021;Xiao, 2021), classification (Liu, Huang, Zhou, & Denoeux, 2021;Liu, Zhang, Niu, & Dezert, 2021) and decision making (Xie, Xiao, & Pedrycz, 2021).…”

Section: Introductionmentioning

confidence: 99%

Maximum Entropy of Random Permutation Set

Deng¹,

Deng²

2022

Preprint

View full text Add to dashboard Cite

Recently, a new type of set, named as random permutation set (RPS), is proposed by considering all the permutations of elements in a certain set. For measuring the uncertainty of RPS, the entropy of RPS is presented. However, the maximum entropy principle of RPS entropy has not been discussed. To address this issue, in this paper, the maximum entropy of RPS is presented. The analytical solution for maximum entropy of RPS and its corresponding PMF condition are respectively proofed and discussed. Numerical examples are used to illustrate the maximum entropy RPS. The results show that the maximum entropy RPS is compatible with the maximum Deng entropy and the maximum Shannon entropy. When the order of the element in the permutation event is ignored, the maximum entropy of RPS will degenerate into the maximum Deng entropy. When each permutation event is limited to containing just one element, the maximum entropy of RPS will degenerate into the maximum Shannon entropy.

show abstract

“…The current scene and the relevant scene are named as the target domain and the source domain, respectively. At present, application fields of transfer leaning in the existing studies can be generally divided into four categories [19]: classification [20], [21], regression [22], [23], feature extraction [24], [25], and clustering [26], [27]. While the first three tasks have been studied quite extensively, research on transfer clustering is very limited despite of the wide range of real-world clustering applications.…”

Section: Introductionmentioning

confidence: 99%

TECM: Transfer Learning-based Evidential C-Means Clustering

Jiao¹,

Wang²,

Liu³

2021

Preprint

Self Cite

View full text Add to dashboard Cite

Clustering is widely used in text analysis, natural language processing, image segmentation, and other data mining fields. As a promising clustering algorithm, the evidential cmeans (ECM) can provide a deeper insight on the data by allowing an object to belong to several subsets of classes, which extends those of hard, fuzzy, and possibilistic clustering. However, as it needs to estimate much more parameters than the other classical partition-based algorithms, it only works well when the available data is sufficient and of good quality. In order to overcome these shortcomings, this paper proposes a transfer evidential c-means (TECM) algorithm, by introducing the strategy of transfer learning. The objective function of TECM is obtained by introducing barycenters in the source domain on the basis of the objective function of ECM, and the iterative optimization strategy is used to solve the objective function. In addition, the TECM can adapt to situation where the number of clusters in the source domain and the target domain is different. The proposed algorithm has been validated on synthetic and real-world datasets. Experimental results demonstrate the effectiveness of TECM in comparison with the original ECM as well as other representative multitask or transfer clustering algorithms.

show abstract

Combination of Transferable Classification With Multisource Domain Adaptation Based on Evidential Reasoning

Cited by 80 publications

References 47 publications

Partial classification in the belief function framework

Partial classification in the belief function framework

Maximum Entropy of Random Permutation Set

TECM: Transfer Learning-based Evidential C-Means Clustering

Contact Info

Product

Resources

About