Did everyone get their naive bayes classifier to work in the end? I'm curious to see what everybody did in terms of feature selection. Tijmen and I opted for a restricted bolzmann machine, saving us some time of boring regex hacking.
Here's our report: MLPR T&T
BTW, could everyone who's reading this blog make him/herself a contributor on this blog? I think it's a good idea to eventually close this blog to outsiders. To make yourself a contributor, go to settings -> permissions and invite yourself.
It looks like a good report. I especially like pages 1 to 10.
ReplyDelete