Affiliations: School of Information Engineering, Wuhan University of Technology, Wuhan, Hubei, PR China
Correspondence:
[*]
Corresponding author. Zhang Chuanchao, School of Information Engineering, Wuhan University of Technology, Wuhan, Hubei, 430070, PR China. E-mail: 574799770@qq.com.
Abstract: In view of the characteristics with big data, high feature dimension, and dynamic for a large-scale intuitionistic fuzzy information systems, this paper integrates intuitionistic fuzzy rough sets and generalized dynamic sampling theory, proposes a generalized attribute reduction algorithm based on similarity relation of intuitionistic fuzzy rough sets and dynamic reduction. It uses dynamic reduction sampling theory to divide a big data set into small data sets and relative positive domain cardinality instead of dependency degree as decision-making condition, and obtains reduction attributes of big intuitionistic fuzzy decision information systems, and achieves the goal of extracting key features and fault diagnosis. The innovation of this paper is that it integrates generalized dynamic reduction and intuitionistic fuzzy rough set, and solves the problem of big data set which cannot be solved by intuitionistic fuzzy rough set. Taking an actual data as an example, the scientificity, rationality and effectiveness of the algorithm are verified from the aspects of stability, diagnostic accuracy, optimization ability and time complexity. Compared with similar algorithms, the advantages of the proposed algorithm for big data processing are confirmed.
Keywords: Intuitionistic fuzzy rough set, similarity relation, relative positive domain, generalized dynamic reduction, large fuzzy decision information system, attribute reduction
