Extracting Event Metadata from Proceedings Titles

Zenodo (CERN European Organization for Nuclear Research)

Abstract

Conferences are considered as one of the core media of scholarly communication in many research disciplines. The articles presented via such opportunities are usually published as proceedings. Generally, the titles of proceedings are composed of a limited subset of English terms that uniquely describe an event. Most proceedings title are available in digital form these days - be it as part of the publishing or when referenced e.g. in citations. Therefore, the titles of proceedings are a potential source for automatically extracting metadata about events. This paper presents a semantic parser to extract conference metadata from English titles of proceedings of scientific events. The parser we present has been designed and tested based on over 43,640 proceedings titles extracted from four complementary, high- quality sources. To improve the flexibility and simplicity of the parser, a combined parser / dictionary based approach is applied. The implementation of the Proceedings Title Parser (PTP) is made available as an open source project at https://github.com/ WolfgangFahl/ProceedingsTitleParser.

Authors 3

1. Wolfgang Fahl Aachen

   RWTH Aachen University

   Affiliation as printed

   RWTH Aachen i5

2. Kai Eckert

   Stuttgart Media University

   Affiliation as printed

   Stuttgart Media University

3. Christoph Lange

   Fraunhofer Institute for Applied Information Technology

   Affiliation as printed

   Fraunhofer FIT