Document Modeling and Clustering using Hypergraph

摘要

Documents have been represented using graphs for many applications like document clustering, document summarization etc. They have also been modeled using the vector space model for various text processing activities. The purpose of this paper is to model text using hypergraph and apply the morphological operator on hypergraph created from the underlying text to get text clusters. The document is considered as a graph and partitioning is applied which finally results in clustering. Here document is modeled as a hypergraph and two methods for text clustering are discussed. The first method uses simple hypergraph and the second method uses a weighted hypergraph. The paper also discusses on how to model multiple documents as hypergraph. The method can be extended for multidocument clustering also.