Figure 4 - uploaded by Bastian Entrup
Content may be subject to copyright.
Source publication
It is assumed that ontologies can be represented and treated as networks and that these networks show properties of so-called complex networks. Just like ontologies “our current pictures of many networks are substantially incomplete” (Clauset et al., 2008, p. 3ff.). For this reason, networks have been analyzed and methods for identifying missing ed...
Contexts in source publication
Context 1
... than one direct parent vertex as can be seen in Fig. 3). This leads to the conditio sine qua non that a tree of n vertices always has exactly n − 1 edges. Another special graph that will be of interest later on is the so-called bipartite graph. A bipartite graph consists also of vertices and edges, but the vertices are of two different kinds (see Fig. 4). An example of such a graph can be a network of co-working in the movie business. Three actors {a, c, e} are linked to the movies {b, d} they appear in. The vertices are of the kind movie and actor. Such affiliation networks are widely used in social network analyses. There are no direct connections between the actors, they are only ...
Context 2
... at the ratio between correctly and incorrectly classified instances, one can see that not every class is predicted equally well (see Table 43). The large red portions of the bars in Fig. 40 show this: The isPartOf relation for example is more often wrongly assigned (10.36%) than used in correct classifications (6.77%). Also the error rate of the author and hometown relation is higher than the correct classifications, and the location relation is not assigned one single time in the test set. This is due to the fact that no ...
Citations
Die Vernetzung von Computern bewirkt die Entstehung eines Netzes aus Texten und, als Folge davon, sozialen Netzen von Nutzern dieser Texte als Schreibern und Lesern. Netzwerke sprachlicher Objekte gab und gibt es zwar auch ohne Digitalisierung und Vernetzung, jedoch weniger umfangreich und wesentlich schwerer, möglicherweise gar nicht in großer Menge analysierbar. Der vorliegende Beitrag befasst sich mit den verschiedenen Typen sprachlicher Netzwerke: Textgeweben, Interaktionsnetzwerken und sozialen Netzwerken. Es werden zentrale Begrifflichkeiten der Netzwerkanalyse erläutert und anhand von Beispielen gezeigt, wie sprachliche Kommunikation auf der Grundlage der Methoden der Netzwerkanalyse aus einer anderen, neuen Perspektive betrachtet werden kann.
