Multi-Granulation Decision-Theoretic Rough Set Based on Maximal Consistent Relation
Science Discovery
Volume 6, Issue 4, August 2018, Pages: 290-297
Received: Aug. 9, 2018; Published: Aug. 10, 2018
Views 156      Downloads 22
Authors
Fan Bingbing, Graduate College Air Force Engineering University, Xi’an, China
Li Jin, Graduate College Air Force Engineering University, Xi’an, China
Chen Xicheng, Graduate College Air Force Engineering University, Xi’an, China
Liu Mengbo, Graduate College Air Force Engineering University, Xi’an, China
Gu Jinghao, Graduate College Air Force Engineering University, Xi’an, China
Liu Ming, Graduate College Air Force Engineering University, Xi’an, China
Article Tools
Follow on us
Abstract
In incomplete information system, combining the advantages of maximal consistent relation and multiparticle theory, this paper proposed the multi-granulation decision–theoretic rough set based on maximal consistent relation based on consistent relation. Firstly, this paper define variable precision maximal consistent relation and dual-variable maximal consistent relation respectively for two kinds of incomplete information systems with different value types. Then, this paper establish optimistic and pessimistic multi-granulation decision–theoretic rough set model by replacing the equivalence relation with the maximal consistent relation in multi-granulation decision-theoretic rough set. Finally, it is proved that the maximum compatible relationship can improve the classification accuracy effectively based on model of optimistic maximal consistent relation, and this paper prove that the robustness of the classification can be improved by multiple classification thresholds at Multi-granulation.
Keywords
Maximal Consistent Relation, Multi-Granulation, Decision–Theoretic Rough Set, Classification Accuracy
To cite this article
Fan Bingbing, Li Jin, Chen Xicheng, Liu Mengbo, Gu Jinghao, Liu Ming, Multi-Granulation Decision-Theoretic Rough Set Based on Maximal Consistent Relation, Science Discovery. Vol. 6, No. 4, 2018, pp. 290-297. doi: 10.11648/j.sd.20180604.20
ADDRESS
Science Publishing Group
548 FASHION AVENUE
NEW YORK, NY 10018
U.S.A.
Tel: (001)347-688-8931