Fuzzy Clustering based on Semantic Body and its Application in Chinese Spam Filtering

E-mail’s text is the main body of an E-mail. Its content is reflected by semantic body formed by a large number of semantic elements, so it is the most authoritative and effective to study semantic body information of spam when analyzing its text. Firstly, this paper takes the advantage of HowNet in analysis of semantic element and analyze semantic bodies in email text, then proposes the method of constructing semantic body and calculation ways of similarity between semantic bodies based on sentence similarity. Secondly, for the problem of Imprecision and Fuzziness existing in current spam filtering technology, we use fuzzy clustering method to solve it. Combining fuzzy clustering with the semantic body, the paper proposes the method of fuzzy clustering based on semantic body. It is different from the traditional methods that semantic body is used as the object to be classified and the similarity between semantic bodies used as similarity coefficient in the proposed method. The method reduces the dimension when we use fuzzy clustering method to deal with text clustering problem. Finally, we apply the new method of fuzzy clustering based on semantic body to spam filtering. The result of the experiment shows that this method is more objective in determining email content when comparing with the method of traditional email filtering in semantic unit. The proposed method reflects much better in recall rate of discernment of email for spam whose meaning is expressed unclearly.