Measuring Preferential Attachment in a Hyper-Textual Dictionary Reference Network: Eksi Sözlük

Existing methods used to analyze the relation between preferential attachment and node degree make use of time dependent measures, which result in limited ability to analyze the temporal characteristics of networks. We introduce time independent measures, which allow us to analyze the networks' preferential attachment behavior in a more precise manner. The two different methodologies are compared on a new complex network data: Eksi Sözlük, which spans the whole lifetime (six years) of a complex network with very precise recordings of the node and edge addition events (i.e. one minute). The relation between the likeliness to receive new links and the present degree of a node is found to be linear. Analyses suggest that time independent measures are better in capturing the dynamics of the network and in some cases, provide results that are very hard to obtain by existing methodologies.