August 31, 2003
Spam words
The Bayesian spam filter Popfile lets you query any word to see the probability that it’s a spam indicator. Some semi-random scores based on the 500+ spams I receive a day:
SPAM INBOX
Sex: 0.53 0.71
Viagra: 0.91 0.0
Behind: -0.10 0.91
Remove: 0.90 0.03
Home: 0.53 0.73
Becoming: 0.0 0.962
Bush: -0.61 0.75
AOL: 0.58 0.68
Nigeria: 0.98 0.0