Chinese Document Clustering Based on Topic Concept Clustering

Nowadays,document clustering technology has been extensively used in text mining,information retrieval systems and etc.The conventional document clustering methods rely on the classical vector-space model using the key words as the feature.However,these methods ignore the semantic relations among the keywords,don′t really address the special problems of document clustering:high dimensionality of the data,and high computation complexity.To solve these problems,based on topic concept clustering,this paper proposes a method for Chinese document clustering.The topic concepts from a document are extracted by using HowNet and are clustered by Chameleon algorithm.The document classification has been performed in terms of the results of topic concept clustering.The document is presented by the concept vector to decrease the dependence relations among of documents.The computation complexity of the document clustering is reduced efficiently and effectively via this method.