Fuzzy Clustering Analysis On Synonymous Adjective In Chinese

Fuzzy Clustering Analysis is a mathematical way of applying the viewpoint and theory in Fuzzy Mathematics and then classifying the objective things by establishing Fuzzy Similar Relations. This thesis embarks from the perspectives of Fuzzy Mathematics, thought that the synonymous relations of the words and expressions in language are precisely one kind of Fuzzy Similar Relations which can be taken as the premise to use Fuzzy Clustering Analysis. The object of our study is synonymous adjective.Since the information carried by characters in the language doesn't identify with the data message in the natural subjects, we should try to find solutions to quantify the linguistic information when carry on Fuzzy Clustering Analysis to the synonymous adjective. The thesis introduces the methods of Similarity Science and plereme analysis in the modern semantics to test the similar degree between two random adjectives in a synonymy paradigm. This similar degree is called similar coefficient, it is necessary for Fuzzy Clustering Analysis and can constitute Fuzzy Similar Matrix. Because Fuzzy Similar Matrix doesn't necessarily have the transitivity, it should be transformed into Fuzzy Equivalent Matrix through the synthesis of its square, then dynamic clustering analysis can be carried on by selecting suitable fiduciary level value. The thesis discusses the question from the following three parts: First, the author makes a brief introduction to the basic concept of Fuzzy Sets, pointing that it provides a powerful means for the research of Fuzzy objects; Second, we focus the discussion mainly on the Fuzzy property of synonymous relations and the specific application of Fuzzy Clustering Analysis to the study of synonym; Third, we choose 10 synonymous adjectives of four kinds in《Synonym Word Forest》to carry on Fuzzy Clustering Analysis. Through the results of Fuzzy Clustering Analysis, we can find that the polymerization of each adjective in one paradigm isn't in the same plane, but can be divided into different levels of subcategories. Fuzzy Clustering Analysis not only reveals the existence of such subcategories, but also identifies the numerical polymerization level at which different words can be put into one category. This enables us to have a more thorough understanding of synonymous phenomenon.