Bug 7162 - Autolearn change from 0.1 to -1.0
Summary: Autolearn change from 0.1 to -1.0
Status: RESOLVED WONTFIX
Alias: None
Product: Spamassassin
Classification: Unclassified
Component: Learner (show other bugs)
Version: SVN Trunk (Latest Devel Version)
Hardware: PC Windows 7
: P2 normal
Target Milestone: 3.4.2
Assignee: SpamAssassin Developer Mailing List
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2015-04-02 16:50 UTC by Kevin A. McGrail
Modified: 2015-04-07 13:25 UTC (History)
2 users (show)


Attachment Type Modified Status Actions Submitter/CLA Status
Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Kevin A. McGrail 2015-04-02 16:50:17 UTC 
On 9/10/2014 9:47 AM, Axb wrote:
> for quite a while I've been playing with autolearn settings
>
> SA's default is:
> bayes_auto_learn_threshold_nonspam    0.1
>
> this *can* cause low scored spam to be learnt as ham.
>
> For several months I've been using
> bayes_auto_learn_threshold_nonspam -1.0
>
> and so far no more false negatives have been learnt as ham which is was hoping for.
> If you're using autolearn, you may want to play with that threshold..

KAM: Based on your expertise with Bayes, should we change the default for 3.4.1? I'm +1 for it.
Comment 1 AXB 2015-04-02 17:32:04 UTC 
(In reply to Kevin A. McGrail from comment #0)
> On 9/10/2014 9:47 AM, Axb wrote:
> > for quite a while I've been playing with autolearn settings
> >
> > SA's default is:
> > bayes_auto_learn_threshold_nonspam    0.1
> >
> > this *can* cause low scored spam to be learnt as ham.
> >
> > For several months I've been using
> > bayes_auto_learn_threshold_nonspam -1.0
> >
> > and so far no more false negatives have been learnt as ham which is was hoping for.
> > If you're using autolearn, you may want to play with that threshold..
> 
> KAM: Based on your expertise with Bayes, should we change the default for
> 3.4.1? I'm +1 for it.

While I still support the idea, I think this shouldn't be changed in a 3.4.x release. Rather wait for 3.5 where ppl will be more inclined to accept more radical changes.
so I'm -1
Comment 2 Kevin A. McGrail 2015-04-02 18:02:48 UTC 
Roger that. Bumping to 3.4.2 to consider again then
Comment 3 Kevin A. McGrail 2015-04-07 13:25:40 UTC 
This has apparently been discussed on bug 6828 before and voted down as well as pointing out historical problems with the same change.