Efficient identification of overlapping communities
ISI'05 Proceedings of the 2005 IEEE international conference on Intelligence and Security Informatics
| Hi-index | 0.00 |
This work experimentally examines different notions of stability of the behavior of individuals and groups in a network of blogs. Our experiments are conducted on data collected from LiveJournal. All stability notions aim to locate stable behavior within an individual's area, which is defined in a variety of manners. Our experiments confirm an earlier observation of the highly dynamic nature of the network. Roughly 70% of the communication of a typical week was not observed in the previous week. Depending on the definition of stability and area used, we find small, but highly stable, sets of individuals with stable behavior in the network.