DYNAMICS OF USER NETWORKS IN ON-LINE ELECTRONIC AUCTIONS

The network of users in the Internet auction site Aukro.cz is analyzed. We show that the fluctuations in the activity of users has non-Gaussian distribution with long-time correlations. The bipartite network of auctions and users has degree distribution following a stretched exponential. The networks observed within a short time windows are highly fragmented and their structure exhibits long memory, which is observed in the power-law decay of correlations, with a small exponent around 0.2.