Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

Semantometrics in Coauthorship Networks: Fulltext-based Approach for Analysing Patterns of Research Collaboration

783 views

Published on

To date, many studies of scientific citation, collaboration and coauthorship networks have focused on the concept of cross-community ties. In this article we explore how Semantometrics can help to characterise the types of research collaboration in scholarly publication networks and the nature of the cross-community ties, and how this information can be utilised in aiding research evaluation. In contrast to the existing research evaluation metrics such as Bibliometrics, Altmetrics or Webometrics, which are based on measuring the number of interactions in the scholarly network, Semantometrics build on the premise that fulltext is needed to understand the value of publications. Using the CORE dataset as a case study, this paper looks at the relation between the semantic distance of authors and their research endogamy value. We identify four potential types of collaboration in a coauthorship network. The results suggest similar measures can be used to provide meaningful information about the nature of collaboration in scholarly publication networks.

Published in: Services
  • Be the first to comment

Semantometrics in Coauthorship Networks: Fulltext-based Approach for Analysing Patterns of Research Collaboration

  1. 1. 1/18 Semantometrics: Fulltext-based measures for analysing research collabora<on Drahomira Herrmannova (@damirah) KMi, The Open University & Petr Knoth (@petrknoth) Mendeley Ltd.
  2. 2. 2/18 Introduc<on •  Up to date many studies of scien<fic cita<on, collabora<on and coauthorship networks have focused on the concept of cross-community -es. •  We explore how Semantometrics can help in understanding the nature of the cross- community <es and in characterising the types of research collabora<on in scholarly publica<on networks.
  3. 3. 3/18 Semantometrics •  A set of metrics for evalua<ng research which build on the premise that fulltext is needed to understand the value of publica<ons
  4. 4. 4/18 Cross-community <es •  Links between communi<es
  5. 5. 5/18 Cross-community <es •  The importance of cross-community <es – In cita<on networks, cross-community cita<on paYerns are characteris<c for high impact papers [Shi. et al., 2010] – Same holds true in case of cross-community scien<fic collabora<on [Newman, 2004; LambioYe and Panzarasa, 2009]
  6. 6. 6/18 How to iden<fy cross-community <es? •  From cita<on/coauthorship network – E.g. betweeness centrality •  From fulltext – Seman<c similarity •  Different types of collabora<on when wri<ng a paper – Emerging vs. established – Interdisciplinary vs. intradisciplinary – Etc.
  7. 7. 7/18 Emerging vs. established research collabora<on •  Endogamy – In social sciences: the prac<ce or tendency of marrying within a social group – In research: collabora<on within a group of authors – Higher endogamy = more frequent collabora<on endo A d A a A d a endo p x L p endo x L p p Publica<on A Set of authors d(A) Papers coauthored by authors in A L(p) Set of all subsets with at least two authors of p
  8. 8. 8/18 Interdisciplinary vs. intradisciplinary research collabora<on •  Seman<c distance of publica<on authors – Higher author distance indicates more distant communi<es – Author publica<on record considered as a single text a dist p A p A p dist a , a a A p ,a A p ,a a p Publica<on A(p) Set of authors of p
  9. 9. 9/18 Types of collabora<on when wri<ng a paper High endogamy Low endogamy High author distance Established interdisciplinary collabora<on Emerging interdisciplinary collabora<on Low author distance Expert group Emerging expert collabora<on
  10. 10. 10/18 Experiment •  What is the distribu<on of the four different types of collabora<on in scholarly literature? •  CORE (core.ac.uk) used as a dataset – Cross-discipline – Enables sampling by authors and ins<tu<ons – Selected sample •  Fulltext documents from Open Research Online repository (ORO) •  All other fulltext publica<ons of the authors from ORO found in CORE
  11. 11. 11/18 Dataset sta<s<cs Fulltext ar<cles from ORO 4,207 Number of authors 8,473 Average number of publica<ons per author 7.61 Max number of publica<ons per author 310 Average number of authors per publica<on 4.31 Max number of authors per publica<on 25 Average number of received cita<ons 0.30 Average number of collaborators 80.23 Total number of publica<ons 30,484
  12. 12. 12/18 Endogamy and author distance distribu<on
  13. 13. 13/18 Endogamy and author distance vs. number of authors
  14. 14. 14/18 Rela<on between author distance and endogamy Established interdisciplinary collabora<on Expert group Emerging interdisciplinary collabora<on Emerging expert collabora<on
  15. 15. 15/18 Types of research collabora<on
  16. 16. 16/18 Types of research collabora<on and “impact”
  17. 17. 17/18 Conclusions •  Fulltext necessary •  Semantometrics are a new class of methods •  We showed one method to recognise types of scholarly collabora<on
  18. 18. 18/18 References •  Xiaolin Shi, Jure Leskovec, and Daniel A Mcfarland (2010) Ci0ng for High Impact •  M. E. J. Newman (2004) Coauthorship networks and pa:erns of scien0fic collabora0on •  R. LambioYe and P. Panzarasa (2009) Communi0es, knowledge crea0on, and informa0on diffusion

×