Finding the right supervisor: expert-finding in a university domain

  • Authors:

  • Fawaz Alarfaj;Udo Kruschwitz;David Hunter;Chris Fox

  • Affiliations:

  • University of Essex, Colchester, UK;University of Essex, Colchester, UK;University of Essex, Colchester, UK;University of Essex, Colchester, UK

  • Venue:
  • NAACL HLT '12 Proceedings of the 2012 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies: Student Research Workshop
  • Year:
  • 2012

Quantified Score

Hi-index 0.00

Visualization

Abstract

Effective knowledge management is a key factor in the development and success of any organisation. Many different methods have been devised to address this need. Applying these methods to identify the experts within an organisation has attracted a lot of attention. We look at one such problem that arises within universities on a daily basis but has attracted little attention in the literature, namely the problem of a searcher who is trying to identify a potential PhD supervisor, or, from the perspective of the university's research office, to allocate a PhD application to a suitable supervisor. We reduce this problem to identifying a ranked list of experts for a given query (representing a research area). We report on experiments to find experts in a university domain using two different methods to extract a ranked list of candidates: a database-driven method and a data-driven method. The first one is based on a fixed list of experts (e.g. all members of academic staff) while the second method is based on automatic Named-Entity Recognition (NER). We use a graded weighting based on proximity between query and candidate name to rank the list of candidates. As a baseline, we use a system that ranks candidates simply based on frequency of occurrence within the top documents.