Automated Citation Matching and Relevancy Score Analysis for Scientific Papers

Acknowledging the scientific contribution of previous research work is necessary to ensure a smooth evolution in scientific fields. Citations play an important role in the ranking of authors, journals, institutions, and organizations. Research papers need to contain appropriate citations to previous works. The identification of inline citations and matching these citations to their corresponding bibliographic references remained an open area of research. The inline citation extraction and matching are performed using citation anchor taxonomy and rule-based method. To determine the relevance of citing and cited articles, three similarity concepts are introduced: metadata-based, citation content, and content-based, calculated using the Bag-of-Words model, cosine similarity, and text similarity methods. The in-text citation identification outperformed the F Score of 97% and the proposed method achieved an NDCG score of 0.78. The results show that the automated citation matching and providing the relevancy of citing and cited papers based on both contents and references.