The lesser or greater of two evils?
Jan. 16th, 2004 03:11 pm![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
lovingbothI use POPFile to help sort my email. It looks at the words used within each email and uses Bayes Theorem to classify it into the categories I've chosen according to what I've taught it.
After a day or two, it achieves over 99.5% accuracy (less than one in two hundred emails needs any corrective action from me) so I was rather surprised when an email offering me 'generic viagra' was in my main inbox just now.
Hmm, look at the header to see what POPFile thinks it is... 'unclassified'??
Open POPFile's control panel to see why... Ah ha!
definitely_spam 0.508370
probable_spam 0.491630
It couldn't decide whether it was 'definitely spam' (deleted on sight by my email program) with slightly more than 50% chance, or just 'probably spam' (stuck in a folder for possible later examination) with slightly less than 50%!
After a day or two, it achieves over 99.5% accuracy (less than one in two hundred emails needs any corrective action from me) so I was rather surprised when an email offering me 'generic viagra' was in my main inbox just now.
Hmm, look at the header to see what POPFile thinks it is... 'unclassified'??
Open POPFile's control panel to see why... Ah ha!
definitely_spam 0.508370
probable_spam 0.491630
It couldn't decide whether it was 'definitely spam' (deleted on sight by my email program) with slightly more than 50% chance, or just 'probably spam' (stuck in a folder for possible later examination) with slightly less than 50%!
