Sunday, October 05, 2008

Ranking universities using polytopes

Courtesy Peter Mucha, here is an article in Science News covering a new arXiv paper by Peter Huggins and Lior Pachter that uses polytopes to examine the US News and World Report rankings of US universities. (Some of you may be aware that Lior Pachter is yet another Techer. He was a Rudd, who I believe graduated in 1994. He's currently on the faculty at UC Berkeley and is one of the pioneers in the new field of "applied algebraic statistics." Actually, algebraic statistics is pretty much a new field of mathematics even before one considers applications.)

One of the ideas I've been bouncing around for several years is the idea of using network science to rank universities. One of these days I'll do that... (My current ranking system project, by the way, involves ranking baseball players using the bipartite graph of pitcher-batter interactions. My collaborators and I don't have any results yet, but stay tuned...)

No comments: