A novel approach for clustering Chinese blogs by embedded sentiment based on graph similarity

Blog clustering is an important approach of web public opinion analysis. In this paper, an integrated graph-based approach for representing and clustering Chinese blogs by embedded sentiment is proposed. Graph-based representation and relevant clustering algorithm are applied. This graph-based blogs representation model considers not only sentiment words but also some structural information. Experimental results show that comparing with applying initial graph-based document representation model and vector space document representation model, the integrated graph-based document representation model has better quality in clustering Chinese blog documents.