Similar to the MLWave article, I just put up an educational post detailing how to make a submission to the Rotten Tomatoes Sentiment Analysis competition using Python (or Excel, :shudder:) and Vowpal Wabbit. I tried some different things and ended up tweaking the VW parameters a bit to get a lower average loss, but I haven't bothered submitting the final predictions to the competition since I was more interested in the education component of the process. Feel free to use it/abuse it/provide feedback however you see fit.
[link][1 comment]