Use of Concept Lattices for Data Tables with Different Types of Attributes
In this paper we describe the application of Formal Concept Analysis (FCA) for analysis of data tables with different types of attributes. FCA represents one of the conceptual data mining methods. The main limitation of FCA in classical case is the exclusive usage of binary attributes. More complex attributes then should be converted into binary tables. In our approach, called Generalized One-Sided Concept Lattices, we provide a method which deal with different types of attributes (e.g., ordinal, nominal, etc.) within one data table. Therefore, this method allows to create same FCA-based output in form of concept lattice with the precise many-valued attributes and the same interpretation of concept hierarchy as in the classical FCA, without the need for specific unified preprocessing of attribute values.
formal concept analysis; concept lattices; data mining; fuzzy logic