The Research On Web Text Clustering Based On Concept Lattice And Rough Set

The formal concept analysis has been greatly applied to many fields, such as software engineering, knowledge discovery and professor Wille presented information retrieval in recent years since it. Concept lattice, which can uncover the relationship between concepts through Hasse Diagram, is the core of formal concept analysis. Rough set theory is raised by Pawlak Z in 1982 is a mathematical tools handling imprecision, vagueness and uncertainty in data analysis and it is a definable subset created by the equivalence classes based on the database and other subsets called upper and lower approximations. It provides a mathematical method of knowledge discovery and has already been widely used in knowledge acquisition, machine learning and other fields. Both formal concept analysis and rough set theory is based on data tables, so they have close contact. Rough set theory use the equivalence relation to classify data tables while concept lattice is based on some data tables with ordering theory, in particular the complete lattice theory to discuss concept layers. This article is mainly research the problem of text clustering based on the concept lattices and rough set. The main body of this thesis includes:(1) This paper advanced the text concept clustering based on the concept lattice.using the concept lattice obtain all concepts from text formal context, and definition the similarity function between concepts. Using the concept to express the text, thus reduced the dimension of term, decreased computation complexity, and obtained a kind of clustering result. (2) This paper advanced the text clustering method based on the rough set model, using the rough set model to express text, within the scope of the term weight change into specifically attribute values,thereby document database change into the decision table. Using the upper approximate measure the maximum roughness between attribute,carry out the text clustering according to the partition ideology.(3) The paper is based on the similarity of concept lattice and rough set and combine the two and put forward a rough concept lattice model of variable precision to handle uncertainty and vague information and defined the approximate mapping of the object set and the attribute set in fuzzy concepts. Two parameterβ₁ andβ₂ showing the roughness between the extent and intent of concepts is introduced. According to the different value of two parameterβ₁ andβ₂ input by user, we can get different fuzzy concept lattice with different roughness, and introduce the variable rough concept lattice model in the application of text clustering...