Training from scratch.

[Robin Bowes]

On Fri, July 16, 2004 7:57, mbox mbarsalou said:
> I would generate these values by looking at the
> bogosity values throughout all my mail and looking at the highest value for ham and the
> lowest value when the computer thinks it is spam.

Wouldn't the bogosity value change as bogofilter is trained? For example, assume you get the same piece of spam three times. The first time it arrives it doesn't get recognised as spam and gets a lowish bogosity value. You manually specify it as spam and it is re-processed. The next time it does get classified as spam and gets a high bogosity score.

Or am I looking at this wrong??

R.
-- 
http://robinbowes.com