I’m afraid this is going to be something of an echo-blog, since I don’t have the paid subscription to the ACM library necessary to allow me to see the original article, but the About Kim weblog has an intriguing and encouraging quote from an academic article on link predictability within social networks:
By running our predictors on some other datasets, we have discovered that performance swells dramatically as the topical focus of the dataset widens. In a narrow field, almost anyone can collaborate with anyone else, and new collaborations are largely random. It would be interesting to make precise a sense in which such new collaborations are simply not predictable from the training data.
From this blurb it looks like there exist conditions within certain social networks in which past collaboration or references are not highly predictive of future collaboration or references. In other words, people make new connections with new people a lot.
I’m going to have to get my hands on this article; it’s too tantalizing.
