Finding relatedness between research papers using similarity and dissimilarity scores

Abstract

Identification of relatedness between research papers is very important in recommender systems, information retrieval, and document categorizing etc. The state-of-the-art approaches compute relatedness using similarity functions by exploiting terms from content and metadata, however, dissimilarity between research papers is not used to compute the relatedness. This research proposes an approach for finding relatedness based on similarity as well as dissimilarity based characteristics of document's content. The relatedness between documents has been computed by combining and normalizing similarity and dissimilarity. It was found that results of our experiment are encouragingly comparable to other content based similarity measuring techniques. © 2014 Springer International Publishing Switzerland.