From jm@jmason.org  Wed Sep 18 12:36:24 2002
Return-Path: <yyyy@example.com>
Delivered-To: yyyy@example.com
Received: by example.com (Postfix, from userid 500)
	id 2B27716F18; Wed, 18 Sep 2002 12:36:24 +0100 (IST)
Received: from example.com (localhost [127.0.0.1])
	by jmason.org (Postfix) with ESMTP
	id 28435F7B1; Wed, 18 Sep 2002 12:36:24 +0100 (IST)
To: "Fox" <lds0062@cdc.net>
Cc: "Justin Mason" <yyyy@example.com>, razor-users@example.sourceforge.net
Subject: Re: [Razor-users] HTML Table - Razor Stats using -lm 4 
In-Reply-To: Message from "Fox" <lds0062@cdc.net> 
   of "Tue, 17 Sep 2002 18:01:40 EDT." <003d01c25e95$ccf547c0$7c640f0a@mfc.corp.mckee.com> 
From: yyyy@example.com (Justin Mason)
X-GPG-Key-Fingerprint: 0A48 2D8B 0B52 A87D 0E8A  6ADD 4137 1B50 6E58 EF0A
X-Habeas-Swe-1: winter into spring
X-Habeas-Swe-2: brightly anticipated
X-Habeas-Swe-3: like Habeas SWE (tm)
X-Habeas-Swe-4: Copyright 2002 Habeas (tm)
X-Habeas-Swe-5: Sender Warranted Email (SWE) (tm). The sender of this
X-Habeas-Swe-6: email in exchange for a license for this Habeas
X-Habeas-Swe-7: warrant mark warrants that this is a Habeas Compliant
X-Habeas-Swe-8: Message (HCM) and not spam.  Please report use of this
X-Habeas-Swe-9: mark in spam to <http://www.habeas.com/report/>.
Date: Wed, 18 Sep 2002 12:36:19 +0100
Sender: yyyy@example.com
Message-Id: <20020918113624.2B27716F18@example.com>
X-Spam-Status: No, hits=-6.4 required=7.0
	tests=AWL,HABEAS_SWE,IN_REP_TO,QUOTED_EMAIL_TEXT
	version=2.50-cvs
X-Spam-Level: 


"Fox" said:

> Before or after I whitelisted all the legit mailing lists that Razor is
> tagging?  I had one false positive in the last four days.  Razor tagged some
> guys person-to-person message because he used an ostrich-in-your-face jpeg
> that is probably pretty popular on the net, and -lm 4 means any single
> attachment in a message that is razored, razors the whole message, if I
> understand it correctly.

Razor folks: is -lm documented anywhere?   BTW, I notice all my *.conf
files in ~/.razor use "lm = 4" by default anyway.

> No, I am not keeping official tally of false positives.  I need to write a
> html interface to do it, and then it would be easy.  I imagine you want
> false positive rate per filter.  I will work on it tomorrow, and maybe in a
> week I will have some stats for false positives.

Yeah, that'd be cool -- much appreciated!   comparing text classifiers
like spam filters, without tracking FPs, is not good.   After all, "cat >
/dev/null" gets a 100% hit rate, but without the FP rate figure of, let's
say 90%, you'd never know it was a bad thing to do ;)

--j.