In this paper we analyze a major part of the research output of the Networked Knowledge Organization Systems (NKOS) community in the period 2000 to 2015 from a network analytical perspective. We focus on the paper output presented at the European NKOS workshops in the last 15 years. Our open dataset, the "NKOS bibliography", includes 14 workshop agendas (ECDL 2000-2010, TPDL 2011-2015) and 4 special issues on NKOS (2001, 2004, 2006 and 2015) which cover 171 papers with 218 distinct authors in total. A focus of the analysis is the visualization of co-authorship networks in this interdisciplinary eld. We used standard network analytic measures like degree and betweenness centrality to describe the co-authorship distribution in our NKOS dataset. We can see in our dataset that 15% (with degree=0) of authors had no co-authorship with others and 53% of them had a maximum of 3 cooperations with other authors. 32% had at least 4 co-authors for all of their papers. The NKOS co-author network in the "NKOS bibliography" is a typical coauthorship network with one relatively large component, many smaller components and many isolated co-authorships or triples.
The European NKOS network has held a long-running series of annual workshops
at the European Conference on Digital Libraries (ECDL), latterly reformed
as the International Conference on Theory and Practice of Digital Libraries
(TPDL). Typically, recent advances of KOS have been reported at the NKOS
workshops, e.g. including the Simple Knowledge Organization System (SKOS)
W3C standard, the ISO 25964 thesauri standard, the CIDOC Conceptual
Reference Model (CRM), Linked Data applications, KOS-based recommender
systems, KOS mapping techniques, KOS registries and metadata, social tagging,
user-centred issues, and many other topics. A comprehensive and well cited
review article on KOS and NKOS topics was published in 2004 [
The motivation of this paper is to analyze the research output of the NKOS community. We are focusing on the informal part of this output, the paper presentations given at the past NKOS workshops (the rst European NKOS workshop in 2000 to the 14th European NKOS workshop in 2015). The specialty of this research output is that these research papers typically are not published in journals or conference proceedings. These papers appear just as oral presentations at the workshop and are documented on the website.
To our knowledge nobody has done an analysis on this part of the NKOS research output before. 2
For our analysis we have compiled an open dataset the "NKOS bibliography"1 which includes 14 workshop programs with all presented papers at ECDL 2000, ECDL 2003-2010 and TPDL 2011-2015 (see Table 1). We added papers from 4 special issues on NKOS which have been edited by members of the NKOS community in the same period (see Table 2). 1 The NKOS workshop bibliography is maintained in the following github repository: https://github.com/PhilippMayr/NKOS-bibliography added journal papers from the four mentioned special issues on NKOS. These journal papers are the only formal publications in our analysis. In the end we manually disambiguated author names of all papers.
Our dataset covers 171 papers in total with a sum of 218 distinct author names. Table 1 provides an overview of all workshop papers2. We can see a slightly decreasing amount of papers presented at the workshops after ECDL has renamed to TPDL in 2011. In average 1.9 authors published a workshop paper. In order to analyze the collaboration of the NKOS community we built a network of all authors at the workshops and special issues and computed the centrality of each author. For this purpose we utilized some standard centrality measures in Pajek3. The network is composed of pairs of author names. Each pair means that two authors cooperated for writing a paper. If we have n papers and the paper i has mi authors, the number of pairs are i=1 n X mi(mi 2 1) if mi 1 (1)
These pairs built the network for our analysis in Pajek. To avoid repetition
of pairs, we gave weight to pairs and this is equal to the number of cooperations
of two authors in di erent papers. Two often used centrality measures of authors
are degree and betweenness. Degree is the number of nodes that a focal node
is connected to and measures the involvement of the node in the network [
To show the quantity of collaboration in the community we measured the degree centrality for each author. Figure 3 shows the percentage of authors with di erent degrees. In this gure we see that 15% (with degree=0) of authors had no co-authorship with others and 53% of them had a maximum of 3 cooperations with other authors. 32% had at least 4 co-authors for all their papers.
Figure 4 shows the authors with high degree (more than 8) in the network.
To detect the in uence of authors on information exchange we calculated the betweenness centrality of authors. Figure 5 indicates the authors with high betweenness (more than 0.001). If we compare betweenness with degree we can see that the ranking of authors has changed. We observe that some authors have lower ranking in betweenness (despite their high cooperation with other authors) in comparison to others with lower degree. Comparing the largest component in Figure 2 with Figure 4 and Figure 5 we see that most authors with top centrality (top degree or top betweenness) are in this component. Nevertheless there are 3 authors ('Eva Mendez', 'Ceri Binding' and 'Antoine Isaac') who are not in the largest component, but are in the list of top authors for betweenness or degree centrality. 5
In this paper we have started to analyze the collaborative research of authors and their connectivity to each other in the special case of European NKOS workshop activities and four special issues on NKOS. The results show most active authors in this community who had an important role in the exchange of information exchange and in connecting researchers. We saw some details of the largest component in this network that covers one third of the authors of the whole network. Also we observed that most authors with highest centrality (degree and betweenness) are in this component. Our analyses show that NKOS workshops were successful in bringing researchers from di erent domains together.
We know that our dataset has some severe limitations. First of all we have included just paper presentations. Editing and organizing activities at the workshops have not been covered in our dataset. This leads to artifacts; e.g. Traugott Koch,4 a long-term organizer of the NKOS workshops and editor of the early JoDI special issues on NKOS, is not included in our dataset and the network.
Second, we have not included the activities of the NKOS community in the United States of America. The website at Kent State5 would be a great starting point to look up more research activities of the US NKOS community.
Third, we have not included bibliometric data to complete our analysis. This
is because most of the NKOS workshop activities (presentations) are not formally
cited or even mentioned in scienti c papers. In di erence to the workshop output,
the few journal papers in the special issues on NKOS are cited. Some works (e.g.
[
We are planning to extend the analysis of the NKOS network. In this way we rst plan to complement the dataset with other NKOS research output. We 4 Traugott Koch was an central protagonist and networker of the European NKOS community. He retired and left the NKOS community in 2012. 5 http://nkos.slis.kent.edu/ also plan to analyze the development of topics in the titles and abstracts of the presentations and papers. Combining network analytic measures with bibliometric analysis (e.g. co-citations, bibliographic coupling) would complement our preliminary observations. We invite people to contribute to the data set.
We thank our colleague Julia Achenbach (GESIS) who helped us to compile and correct the dataset. We thank Peter Mutschke (GESIS) who helped us with the interpretation of the NKOS co-authorship network.