UserPreferences

TrainToExhaustion


Note: This wiki is now frozen; you can no longer edit it, and no interactive features work.

1. Train To Exhaustion

While this sounds like a gruesome workout technique, it seems to be a useful training regimen. Basically, you score pairs of ham and spam, only training on those which don't score within the proper range (below the ham cutoff or above the spam cutoff). Gary Robinson has a [http://www.garyrobinson.net/2004/02/spam_filtering_.html good description] of the technique on his website. There's a tte.py script in the contrib directory of the Spambayes distribution which I wrote based upon that essay. I've been using it for a few months and am quite happy with it.

SkipMontanaro