Title: Mining description logic concepts from relations based on concept lattices
Authors: Yuxia Lei; Feng Jiang; Juan Chen; Yonghua Han
Addresses: School of Information Science and Engineering, Qufu Normal University, Yantai Street 80, Rizhao, 276826, China ' School of Information Science and Technology, Qingdao University Science and Technology, Qingdao, 266042, China ' School of Information Science and Engineering, Qufu Normal University, Yantai Street 80, Rizhao, 276826, China ' School of Information Science and Engineering, Qufu Normal University, Yantai Street 80, Rizhao, 276826, China
Abstract: In description logics (DLs), a concept C could be a set of some tuples or a subset of the domain of an attribute. This paper proposed a method for extracting descriptive concepts from a relation, which is based on formal concept analysis (FCA). FCA can be used to support the descriptive concept construction in domains of attributes. Our method can be described as follows: given a relation R and some conceptual partitions on domains of some attributes, by classify-based scaling, one can obtain a new relation R, and correspondingly obtain its concept lattice in which all concept extents are regarded as concepts in a set of tuples. By using these concepts in tuples, one can further construct concepts in domains of attributes. Conversely, one can further construct concepts in tuples. The paper also defined fuzzy concepts, and discussed the connections between fuzzy concepts and those in fuzzy description logics.
Keywords: relational databases; concept lattices; description logic concepts; fuzzy concepts; data mining; knowledge discovery; concept lattices.
International Journal of Collaborative Intelligence, 2016 Vol.1 No.3, pp.222 - 236
Received: 24 Jun 2015
Accepted: 02 Jul 2015
Published online: 21 Jun 2016 *