An analysis of the dynamics of a social network

The internet community pussokram.com can be considered as a social network, which developed in time. Data on times when contacts between members took or received are available. We will use two process models to study the dynamic properties of the network. A pure birth process model is used to describe the time between contacts. A probability model for the order, which new contacts are established is also applied. In the analysis, ML-estimates and profile log-likelihood confidence intervals are calculated. ∗Postal address: Mathematical Statistics, Stockholm University, SE-106 91, Sweden