You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@spamassassin.apache.org by John Fleming <jo...@wa9als.com> on 2004/02/08 10:59:32 UTC

How to start over with Bayes

I didn't get an answer to my first post, so I want to try again.

I want to know how to "start fresh" with Spamassassin/Bayes.  I want to know
this in general, and I have a specific reason too.

My mail is in mbox format, but the first time I ran sa-learn, I didn't
specify the --mbox option.  Thus sa-learn thought it was processing one HUGE
email that was actually 400+ mails all in a long file, including headers
etc.  Thus I assume a lot of stuff was evaluated that shouldn't have been?

I deleted all the user-specific files I could find with bayes in them, eg
bayes_toks.  Then I uninstalled and reinstalled Spamassassin.  However, when
I ran sa-learn on the spam corpus with the proper --mbox option, it said it
learned from zero messages, although it correctly reported that it examined
400+ - Thus it remembered seeing those emails from previous, and I still
don't have a fresh slate.

Maybe it's OK as I did it originally, but I still want to know how to start
completely over fresh and will appreciate that specific information.  Please
don't tell me to run --forget - I want to start completely over.  Thanks -
John



Re: How to start over with Bayes

Posted by John Fleming <jo...@wa9als.com>.
----- Original Message ----- 
From: "Alan Baxter" <at...@baxtersys.com>
To: "John Fleming" <jo...@wa9als.com>
Cc: <sp...@incubator.apache.org>
Sent: Sunday, February 08, 2004 10:47 AM
Subject: Re: How to start over with Bayes


> On Sun, 8 Feb 2004 04:59:32 -0500, "John Fleming" <jo...@wa9als.com>
> wrote:
> >Maybe it's OK as I did it originally, but I still want to know how to
start
> >completely over fresh and will appreciate that specific information.
Please
> >don't tell me to run --forget - I want to start completely over.
 Thanks -
> >John
>
> Delete bayes_seen, bayes_toks, and bayes_journal.

THANK YOU THANK YOU - I missed the journal!



Re: How to start over with Bayes

Posted by Alan Baxter <at...@baxtersys.com>.
On Sun, 8 Feb 2004 04:59:32 -0500, "John Fleming" <jo...@wa9als.com>
wrote:
>Maybe it's OK as I did it originally, but I still want to know how to start
>completely over fresh and will appreciate that specific information.  Please
>don't tell me to run --forget - I want to start completely over.  Thanks -
>John

Delete bayes_seen, bayes_toks, and bayes_journal.

Alan
-- 
Please avoid sending me Word or PowerPoint attachments.
See http://www.fsf.org/philosophy/no-word-attachments.html

Re: How to start over with Bayes

Posted by Tim B <mo...@optonline.net>.
John Fleming wrote:

> I didn't get an answer to my first post, so I want to try again.
> 
> I want to know how to "start fresh" with Spamassassin/Bayes.  I want to know
> this in general, and I have a specific reason too.

AFAIK, you should be able to just delete the bayes database files.