Research On Constructing Method Of Iceberg Concept Lattice And Fuzzy Concept Similarity

Formal concept analysis(FCA) is a powerful tool for conceptual knowledgeprocessing. It has to be used and developed in many fields, such as informationretrieval, data mining, software engineering and knowledge discovery. With the rapidincrease of information and the size of the database, the database contains not onlyuseful information, but also a lot of redundancy. It has practical value that we deriveuseful informationn from the database.For the large databases, because of generating large-scale concept lattice andanalyzing difficultly, so that it lose its value. Iceberg concept lattice is a conceptualclustering method, it is suitable for the analysis of large databases and reflect the criticalinformation in the database. But the semi-lattice structure of iceberg concept latticestructure, which makes perfect concept lattice algorithm can not adapt to the icebergconcept lattice structure, so it is very important significance to study iceberg conceptlattice construction.Besides concept similar measure based on concept lattice is used and maped thetraditional formal ontology integration. However, many application areas and most ofthe information is complex and uncertain, it can not express these vague and uncertaininformation problems with the traditional formal concept analysis. Therefore, studyingthe concept of similarity in the formal background is very necessary.Iceberg concept lattice structure is a prerequisite for their applications. There arethree points in this paper: firstly, we propose merger iceberg concept lattices andconstruct Algorithm, we construct sub iceberg concept lattices with spliting in the formof background. Then we merge the sub iceberg concept lattices to get completeiceberg concept lattices. Secondly, we put forward the method of iceberg concept latticein the fuzzy formal background. Thirdly, we improve the similarity measure methodbased on fuzzy concept lattice structure, the traditional measured introduce the distanceof concepts, the conceptual level and structural factors by adjusting the weights to suitdifferent needs.