BibMix: Enrichment of Citation Metadata Based on Integration of Bibliographic Data

A citation in an academic text provides the reader with exact publication information to uniquely identify and locate a scholarly resource for further study. In the domain of academic research, citations are a key part of linking discrete pieces of knowledge. As opposed to the natural language mostly used in academic texts, bibliographic applications are typically based on structured publication information and therefore usually rely on the quality and completeness of automatically extracted citation metadata. However, the results from automatic Citation Metadata Extraction approaches are often unreliable. Moreover, the extracted citation metadata is only a small part of the relevant metadata that is available spread across the internet. In this thesis, we propose an approach and a supporting software component capable of enriching citation metadata with relevant bibliographic data from sources on the web. For the evaluation of the approach, we have designed a quality measure to express the quality of enriched references as a number. Based on this measure we have assessed the quality of the enriched references of BibMix to be 0.72. Graduation Committee: Prof. dr. ir. G.J. Houben, Faculty EEMCS, TU Delft Dr. Laura Hollink Dr. Martin Pinzger Samur Araujo M.Sc.