Citation Clustering for Identifying Research Contribution

Abstract

The h-index is an index that measures productivity and citation impact of the published work but it has been criticized because it does not consider context of citation and reason behind citation. This indicates that there is a need for an improved h-index by a new approach which includes important citations received by a paper instead of the whole list of citations. Citation classification is an emerging area of research that categorizes citations based on the purpose behind the citation. To perform citation classification there is need of a standard set of classes called as classification scheme. Such standard scheme is not available so we aim to generate a citation classification scheme automatically i.e. by using hierarchical clustering. The clustering is performed by using similarity vectors. The main contribution of this research is to generate similarity distance matrix of keywords and verbs extracted from the citation sentences with the help of WordNet.