A frequent keyword-set based algorithm for topic modeling and clustering of research papers

摘要: In this paper we introduce a novel and efficient approach to detect topics in large corpus of research papers. With rapidly growing size academic literature, the problem topic detection has become very challenging task. We present unique that uses closed frequent keyword-set form topics. Our also provides natural method cluster papers into hierarchical, overlapping clusters using as similarity measure. To rank cluster, devise modified PageRank algorithm assigns an authoritative score each by considering sub-graph which appears. test our algorithms on DBLP dataset experimentally show are fast, effective scalable.