The main publishers in Computer Science

Geplaatst op 27-02-2012 door Maarten Marx | data | tags: | comment image Geen reacties »

With the current protest against Elsevier, PoliticalMashup decided to try to find out how important Elsevier is in Computer Science.
Our research shows that Elsevier is very important in CS. It is the third largest publisher, publishing 13.5% of all CS publications.

How was this computed?
We downloaded a dump from DBLP and extracted all links to electronic editions of journals. From the almost 3M articles in DBLP, 1.65M have such a link (specified in the EE field).
These links are digital object identifiers (DOI). These DOI’s have a rather strict structure. In particular, every publisher has a unique code of the form 10.d{4}. So for each such code, we counted how many different DOI’s with that code are referenced in DBLP.

We found 255 different publishers in DBLP with at least two publications. The DOI site does not provide a way of resolving publishers, so we did the top 5 by hand.
Here is the top 10. The first column contains the number of electronic publications in DBLP by that publisher, the second column contains the DOI code of the publisher.

 438851 10.1109  IEEE
 375702 10.1007  Springer
 223711 10.1016  Elsevier
 181429 10.1145  ACM
  24361 10.1093  Oxford University Press
  19027 10.1002
  15178 10.1023
  13695 10.1142
  12527 10.1137
  12368 10.1080

As expected, this is a rich-get-richer distribution. The first publisher, IEEE, has over 25% of all links. The first two have half of all links, and the first four (less than 2% of all publishers) have roughly three quarters of all publications. The fifth largest publisher (OUP) has less than 1.5% of all publications.


Je moet ingelogd zijn om te kunnen reageren.