If you're new to NLP, I recommend getting a book like Natural Language Processing with Python, using the Python Twitter API, and writing a Bayesian spam classifier. If you're less new, I've been working in sentiment classification for a while now and it's a lot of fun. Also things like automatically grouping tweeters by topic, or scanning someone's tweets to see if you should follow them, can be cool projects.
j On Sun, Feb 28, 2010 at 10:16 PM, Bryan <k0...@dbbear.com> wrote: > Google: Latent Semantic Analysis, Latent Relational Analysis, and > Vector Space Model > > Check out the book: Algorithms of the Intelligent Web from Manning: > http://www.manning.com/marmanis/ > > I've blogged a bit about some of the things I applied to Twitter at: > http://skimmeragent.blogspot.com/ > > Good luck. > > Bryan >