Open AccessAttribute reduction based on improving DIT in interval‐valued ordered information system
Author(s) -
Yang Lei,
Zhang Xiaoyan,
Xu Weihua,
Sang Binbin
Publication year - 2020
Publication title -
the journal of engineering
Language(s) - English
Resource type - Journals
ISSN - 2051-3305
DOI - 10.1049/joe.2019.1158
Subject(s) - reduction (mathematics) , equivalence relation , mathematics , ranking (information retrieval) , matrix (chemical analysis) , interval (graph theory) , equivalence (formal languages) , relation (database) , data mining , algorithm , computer science , artificial intelligence , discrete mathematics , combinatorics , materials science , geometry , composite material
Traditional attribute reduction based on discernibility information tree (DIT) is proposed under the equivalence relation, which effectively realises the compression and storage of discernible matrix. However, it has not been studied whether it is applicable under the dominant relation. In this study, under the background of an interval‐valued ordered information system, the improved DIT (IDIT) based on a discernible matrix is constructed by combining the core attributes and attribute significance. Furthermore, the compressed storage of the discernible matrix is realised. Moreover, only one reduction of the information system can be found through traditional DIT. By changing the ranking order of attributes in the information system, this study finds all reductions in the interval‐valued ordered information system by the IDIT.