High accuracy citation extraction and named entity recognition for a heterogeneous corpus of academic papers

Citation indices are increasingly being used not only as navigational tools for re- searchers, but also as the basis for mea- surement of academic performance and re- search impact. This means that the reliabil- ity of tools used to extract citations and con- struct such indices is becoming more crit- ical; however, existing approaches to cita- tion extraction still fall short of the high ac- curacy required if critical assessments are to be based on them. In this paper, we present techniques for high accuracy extrac- tion of citations from academic papers, de- signed for applicability across a broad range of disciplines and document styles. We in- tegrate citation extraction, reference pars- ing, and author named entity recognition to significantly improve performance in cita- tion extraction, and demonstrate this per- formance on a cross-disciplinary heteroge- neous corpus. Applying our algorithm to previously unseen documents, we demon- strate high F-measure performance of 0.98 for author named entity recognition and 0.97 for citation extraction.