Dispersion Based Similarity for Mining Similar Papers in Citation Network

Measuring "similarity" has been established as afundamental problem and has been widely studied. In thispaper we propose a novel approach for establishing similarityin context of citation network. With the rapidly growing sizeof academic literature, the problem of finding similar researchpapers has become a challenging task. Research papers in acitation network often form communities based on an underlyingconcept. Our research shows that dispersion based similaritymeasure can be used as a strong measure for finding similarpapers based on similar connectivity in those communities andstructural relevance of the citation network. Our results showthat our approach works better than other conventional link-based similarity measures both quantitatively and qualitatively. One of the direct benefits of this research is to support the highlyspecialized information needs of a scholarly researcher workingin a specialized field of research.