Explores a method for pruning co-occurrence networks (e.g., co-citation). Presented at STI 2014.

1. Pruning cooccurrence networks
Raf Guns
raf.guns@uantwerpen.be

2. Bibliometric networks

3. Density problem
Dense networks are hard to
 visualize
 interpret
Solution: pruning networks
 PathFinder (Schvaneveldt, 1990)
 Deleting low-weight links (De Nooy, Mrvar, and Batagelj, 2005)
 Cocitation and bibliographic coupling (Persson, 2010)
 Threshold for cosine values (Leydesdorff, 2007; Egghe &
Leydesdorff, 2009)

4. Cooccurrence networks
E.g. cocitation, bibliographic coupling, coauthorship…
Especially prone to density problem
Two-mode network Cooccurrence network
e.g., authors
e.g., citing
papers

5. Methods

6. Steps
Based on Zweig and Kaufman (2011): we start from two-mode
network
1. Define pattern of interest
2. Determine interestingness of cooccurrence
3. If cooccurrence is interesting, authors are linked

7. Why interestingness?
Highly cited author
 High coocurrence counts with many other authors
Citing paper referring to many authors under consideration
 Resulting cooccurrences are less important

8. Determining interestingness
Here:
How to determine Exp and σ?
 Estimate by sampling from Fixed Degree Sequence Model
(FDSM): all two-mode networks with same node degrees
 Markov Chain Monte Carlo simulation: link swapping
 If p < 0.0001 (or z > 3.29) , we consider link interesting

9. Link swapping
e.g., authors
e.g., citing
papers

10. Link swapping

11. Results

12. Author cocitation
Author (co-)citations to
 12 authors from bibliometrics
 12 authors from information retrieval
in Scientometrics and JASIS, 1996-2000
Same data set studied by
 Ahlgren, Jarneving & Rousseau (2003)
 Egghe & Leydesdorff (2009)
 Leydesdorff & Vaughan (2006)

13. Author cocitations: cosine

14. Author cocitations: FDSM and z-scores

15. Bibliographic coupling
Bibliographic coupling of all JASIST articles, 1999-2000
 n = 371
 12 981 unique references
Two VOSviewer maps
 cosine normalization
 FDSM and z-scores

16. Bibliographic coupling: cosine

17. Bibliographic coupling: FDSM and z-scores

18. Conclusions
Advantages
1. Both positive and negative cooccurrences
2. Thresholds correspond to specific p-values
3. Accounts for degree variations of bottom nodes
Disadvantages
1. Some nodes may become isolates
2. More computationally intensive than cosine similarity

19. Thank you!