Journal
INTERNATIONAL JOURNAL OF APPROXIMATE REASONING
Volume 140, Issue -, Pages 7-30Publisher
ELSEVIER SCIENCE INC
DOI: 10.1016/j.ijar.2021.09.017
Keywords
Multi-granulation; Information fusion; Decision-theoretic rough sets; Granular computing
Categories
Ask authors/readers for more resources
The paper introduces an Adaptive Multi-Granulation Decision-Theoretic Rough Sets (AMG-DTRS) model, which adaptively obtains probabilistic thresholds by setting a compensation coefficient. Three types of mean AMG-DTRS models are studied, offering a new perspective on information fusion. The advantages and generalization of the AMG-DTRS model are demonstrated by analyzing its connections and differences with existing MGRS models.
The Multi-Granulation Decision-Theoretic Rough Set (MG-DTRS) is an effective method for cost-sensitive decision making from multi-view and multi-level. However, the inherent weak point of MG-DTRS model is to compute three regions with a subjectively given pair of probabilistic parameters (i.e., alpha and beta). To overcome this issue, this paper first proposes a generalized MG-DTRS model called Adaptive Multi-Granulation Decision-Theoretic Rough Sets (AMG-DTRS), which can adaptively obtain a pair of probabilistic thresholds by setting a compensation coefficient zeta. Then, three types of mean AMG-DTRS models are investigated, which provide a novel perspective on multi-granulation method for information fusion. Finally, the connections and differences between the proposed AMG-DTRS and the existing MGRS models are analyzed, which show the advantages and generalization of the AMG-DTRS model. In addition, there are numerous existing MGRS models can be derived explicitly by considering various MG-DTRS, MGRS and VP-MGRS based on our model. These results will be conducive to establishing the framework of information fusion for granular computing. (c) 2021 Elsevier Inc. All rights reserved.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available