August 13, 2004
At the beginning of the week I finished the report “Discovering Communities through Information Structure and Dynamics,” (pdf) a review of some recent research that provides insight into how the existence of, and structure of, computer-mediated communications can be used to learn about online communities. I gave a presentation on this topic on Monday, finishing my last preliminary exam. With my other requirements done, I now have only my dissertation to complete — of course, I need to begin it first, but if I can manage two years of focused research work, that could earn me a Ph.D.
The report looks in some depth at three studies of automated community characterization and identification online: “Inferring Web Communities from Link Topology” (pdf) by David Gibson, Jon Kleinberg, and Prabhakar Ragavan; Email as Spectroscopy: Automated Discovery of Community Structure within Organizations (pdf) by Joshua Tyler, Dennis Wilkinson, and Bernardo Huberman; and “Dialog in E-mail Traffic” (pdf) by Jean-Pierre Eckmann and Elisha Moses. Two other studies are considered in a bit less depth, and Kleinberg’s paper about Hypertext-Induced Topic Search (HITS), the basis for the Gibson, Kleinberg, and Ragavan work, is discussed in some detail, too.