September 28th, 2004

  • evan

social networks research

Finding community structure in very large networks is a better implementation of community-finding algorithm I worked on a while back. They claim the normal algorithm is O(n^3), while theirs nears O(n) for networks with certain properties.

The researcher I was talking to a long while ago published his Statistical Analysis of the LiveJournal Network, though it's from ancient data. And somehow Brad and I ended up as Fitzgerald and Martine, even though he got Pete Krawczyk/sachmet's name right?

cemcom is looking to do some research, too. Maybe brad will see this post and take pity on them. :)
  • evan


Yahoo's search blog discusses blogging services. The graph is "which service are people searching for on yahoo?", which is a weird metric, but it might correlate to, uh, how dumb the users are-- how often do you search for a service rather than just typing in its address?
In a comment, the poster adds, "The teen bloggers on Xanga and LiveJournal right now might be the adopters of Blogger and Typepad down the line. Right now, searches on Blogger are about one-tenth of Live Journal and one-fifteenth of Xanga. Typepad doesn't have enough searches to enter the fray right now."

But I'm especially amused by the comment by "Andrew", which claims LJ ain't all bad 'cause people like jwz use it. Heh.