A Novel Document Representation Model for Clustering

Text document plays an important role in providing better document retrieval, document browsing and text mining. Traditionally, clustering techniques do not consider the semantics relationships between words, such as synonymy and hypernymy. Existing clustering techniques are based on the syntactic structure of the document. To exploit semantic relationships, WordNet has been used to improve clustering results. However, WordNet-based clustering methods mostly rely on single-term analysis of text; they do not perform any phrase-based analysis. To address these issues, we derive the semantic structure of the document. Case grammar structures from the field of natural language processing, are used as semantic structure. These structures are used as document representation model and used for clustering. Semantic similarity measure is used to compare the documents’ similarity. The experimental results show the effectiveness of semantic relationships for clustering. Quality of the cluster has been improved. Moreover, semantic structure improves the WordNet-based clustering method.