Berlin 2012 – scientific programme

Parts | Days | Selection | Search | Updates | Downloads | Help

DY: Fachverband Dynamik und Statistische Physik

DY 18: Networks II (with SOE)

DY 18.9: Talk

Wednesday, March 28, 2012, 12:30–12:45, H 0110

A comparison of probabilistic distribution for fitting the degree distribution of real-world social networks — •Faustino Prieto and Jose Maria Sarabia — University of Cantabria, Santander, Spain

In this paper, the degree distribution of social and information networks is analyzed. Several functional forms have been proposed in the network science literature, including the classical power law and many others. Now, six different probabilistic models are fitted in the entire range by maximum likelihood. The used models are Lognormal, Gamma, Weibull, power law and two special cases of the Pareto Positive Stable (PPS) distribution. The models are compared using the Akaike information criterion (AIC) and the Kolmorov-Smirnov (KS) statistic. A two-parameter PPS distribution is found to be the better choice in the whole range, to several social and information network datasets. Finally, the PPS model is validated graphically by using log-log rank-size plots and double log-log plots.