Wednesday, October 06, 2010

More Tales from Statistics Class

Statistics class was cancelled today because fall break is tomorrow and Friday, and I'm pretty sure that he didn't feel like teaching. My hunch is that teaching this class is nowhere on his priority list, and he is not doing much in terms of planning. The slides are all recycled -- some from last year, some from a talk he gave at a conference last year. The material from his talk is actually pretty cool. Not presented well (wordy slides read verbatim), but they were mathematically pretty interesting. Didn't make complete sense, but seemingly interesting.

So today I went in to ask questions about what he'd been talking about. I kept asking questions and follow-up questions and follow-ups on the follow-ups until I finally figured out that this method is pretty much a breadth-first-search on a Markov chain.

He gave me two (very large) papers to read about his method (from his book).

Really what I want to do is write some code that illustrates the progression of the algorithm. I want to draw every point in the space and watch the system progress as the algorithm runs. The trick is how to make a picture that makes sense with thousands of points.