Kmail and spam filtering

[Nigel Henry]

On Friday 28 July 2006 01:14, Thierry de Coulon wrote:
> On Friday 28 July 2006 00.05, Nigel Henry wrote:
> > Hi Thierry.  I've got Bogofilter 1.0.2 working ok. It's separating the
> > Spam, and the Ham ok, although I've had a few problems setting it up.
> > There appear to be quite a few changes in 1.0.2 from the version you're
> > using.
>
> (...)
>
> > Also I havn't figured out how to get mail that bogofilter is "Unsure"
> > about set up.  Adding a third line to "Filter Criteria" for "Unsure", and
> > then applying, just results in the line being greyed out.  Apart from
> > that, I'm getting very good results from bogofilter. No false negatives,
> > and only 2 false positives up to now.
> >
> > Nigel.
>
> Hi Nigel,
>
> Thanks for all those information. I'm compiling a new Gentoo installation
> just now (though I should actually go to bed...). I'll see what version of
> Bogofilter I get and fiddle with those settings.
>
> I have to think about "unsure" - it does not make a lot of sense to me, as
> "unsure" has to be checked so it belongs in "normal" mail.
>
> What do you call "false positives"? To me, false positives would be "ham"
> sorted out as "spam"

Hi Thierry. Yes that's what I mean. I'm getting the odd bit of Ham turning up 
in the spam directory.
>
> False negatives (spam sorted out as ham) cannot be prevented, they usually
> use "new" spam solutions. What I can't accept are correct mails that are
> considered to be spam - because that means that if my spam goes directly to
> he trash I'll loose mails. I agree that people shoud *not* send html mails,
> be I have friends you won't bring to understand this, and probably you have
> too...

Yes, I'm keeping my spam directory, rather than sending the Spam to the trash, 
just in case some legit mail ends up there. Also I can keep building up the 
wordlist.db (thats another change. spamlist.db, and goodlist.db, have been 
replaced by wordlist.db).  I think that is what "Unsure" is supposed to do. 
Bogofilter looks at the mail. It looks like it might be spam, but isn't sure. 
So rather than sending it to the trash, it sends it to the unsure folder, 
where you can have alook at it.  Possibly the 2 bits of Ham that I've had 
wrongly identified as Spam, would have ended up in the unsure box, if I had 
figured out how to set it up. I think the same works with Spam. If bogofilter 
isn't sure if it's Spam, it again puts it in the unsure folder.
>
> My version of bogofilter misses most html mails, so I had to set-up further
> kmail filters to catch these, but I put them in a special directory and
> check them because many would be "false positives".
>
> Anyway, I prefer a solution of "no provider filtering" + local Bogofilter
> as that solution is both reliable and under my control.

Agreed.

I must say I'm very impressed with it. Apart those 2 false positives, all of 
the Spam has ended up in the spam directory, and all the Ham, in the inbox.

I think I will put the question about "Unsure" on the Fedora mailing list, and 
perhaps the LQ forum. Perhaps someone is using version 1.0.2 with Kmail, and 
may some more up to date info, especially as most of the info on Google is 
pretty ancient.
>
> Thierry

Nigel.
___________________________________________________
This message is from the kde mailing list.
Account management:  https://mail.kde.org/mailman/listinfo/kde.
Archives: http://lists.kde.org/.
More info: http://www.kde.org/faq.html.