good.mbx

开源项目BOGOFILTER的源码.该项目实现BAYES文本分类算法,不过该源码只支持英文的分词

From mdlug-return-6174-smythe=[e-mail]  Fri Nov  1 05:42:47 2002
Return-Path: <mdlug-return-6174-smythe=[e-mail]>
Delivered-To: [e-mail]
Received: from mdlug.org (cpe-24-221-90-239.mi.sprintbbd.net [242.21.90.239])
	by example.com (Postfix) with SMTP id ADE5027F82
	for <[e-mail]>; Fri,  1 Nov 2002 05:42:46 -0500 (EST)
Received: (qmail 21544 invoked by uid 504); 1 Nov 2002 10:38:57 -0000
Mailing-List: contact [e-mail]; run by ezmlm
Precedence: bulk
X-No-Archive: yes
list-help: <mailto:[e-mail]>
list-unsubscribe: <mailto:[e-mail]>
list-post: <mailto:[e-mail]>
Reply-To: [e-mail]
Delivered-To: mailing list [e-mail]
Received: (qmail 21532 invoked from network); 1 Nov 2002 10:38:57 -0000
Message-ID: <[e-mail]>
Date: Fri, 01 Nov 2002 05:38:30 -0500
From: Tom Amazon <[e-mail]>
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.0.0) Gecko/20020623 Debian/1.0.0-0.woody.1
MIME-Version: 1.0
To: [e-mail]
References: <[e-mail]>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [MDLUG] Still sleeping with the gorilla?


I shouldn't, but I must....

Bill wrote:
> What happens when the gorilla has a bad dream?
> 
> http://www.silicon.com/bin/bladerunner?REQUNIQ=1036136703&30REQEVENT=&REQINT1=56175&REQINT2=1&REQSTR1=Text%20News&REQSTR2=Microsoft%20stirs%20up%20Office%2011%20support%20storm&REQSTR3=%7c&REQAUTH=21046&14006REQSUB=
> 
> "A representative acknowledged that Microsoft plans to offer the productivity 
> suite, code-named Office 11, only for Windows 2000 with Service Pack 3 and 
> Windows XP.
> 
> By focusing on development for the single Windows code base, Microsoft could 
> conceivably deliver a better version of Office and cut development costs in 
> the process. But the company also has a tendency of being heavy-handed with 
> customers who resist making the transition to a newer operating system.
> 
As stated, this is nothing new.  This is an extremely important 
fact to remember.

> 
> How, for instance, are you going to be able to justify breaking up Microsoft 
> if the company only has one (kitchen sink dept.) office product running on 
> one OS? That's what the MSIE fuss was all about. 
>
What makes you think that anything like this will ever happen?
Consider this.  By the time anything is actually decided upon, 
Microsoft will no longer support a single product that was in 
existence at the time of accusation.  Therefore there will be no 
real activities to undertake and Microsoft will continue with 
business as usual.  If you can't move the gun, move the moose.

> What happens when the office suite becomes part of the OS, too?
> 
It already is more than you want to know.

> 
> This game aint over. Not by a long shot.
> 

Nope.
But's it's changing.  See Important Fact number one.
People are starting to disbelieve Microsoft more and more.

People are starting to hear whispers about another one.  Remember 
Yoda?  "There is another..."

People are starting to reconsider the finances of owning a 
computer.  They are starting to learn that upgrading is no 
guarantee of function.

Companies are the slowest.  If a really big corporation (auto 
industry) announced that everyone was going to open office and to 
deal with it, then there would be a groundswell in activity.
Why they don't is beyond me.  They dictate CAD/CAM/CAE 
software/hardware all the time.  Why would the rest be any different?
-- 
We question most of the mantras around here periodically, in case
you hadn't noticed.  :-)
              -- Larry Wall in <[e-mail]>

From [e-mail]  Fri Nov  1 06:20:26 2002
Return-Path: <[e-mail]>
Delivered-To: [e-mail]
Received: from csl2.contronics.on.ca (csl2.contronics.on.ca [240.138.93.2])
	by example.com (Postfix) with ESMTP id F007927F82
	for <[e-mail]>; Fri,  1 Nov 2002 06:20:25 -0500 (EST)
Received: from athame.dynagen.internal(localhost[127.0.0.1]) (2517 bytes) by csl2.contronics.on.ca
	via sendmail with P:esmtp/R:bind_hosts/T:inet_zone_bind_smtp
	(sender: <[e-mail]>) 
	id <[e-mail]>
	for <[e-mail]>; Fri, 1 Nov 2002 06:20:25 -0500 (EST)
	(Smail-3.2.0.114 2001-Aug-6 #5 built 2002-Mar-26)
Received: (2148 bytes) by athame.dynagen.internal
	via sendmail with P:stdio/R:smart_host/T:smtp
	(sender: <[e-mail]>) 
	id <[e-mail]>
	for [e-mail]; Fri, 1 Nov 2002 06:20:24 -0500 (EST)
	(Smail-3.2.0.111 2000-Feb-17 #5 built 2000-Apr-12)
Date: Fri, 1 Nov 2002 06:20:24 -0500
From: Greg Lamb <[e-mail]>
To: David Smythe <[e-mail]>
Subject: Re: writeup revision
Message-ID: <[e-mail]>
Reply-To: Greg Lamb <[e-mail]>
References: <[e-mail]> <[e-mail]> <[e-mail]>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
In-Reply-To: <[e-mail]>
Organization: Dynagen Consulting Limited

On 20021031 (Thu) at 1805:18 -0500, David Smythe wrote:

> Also, I'm thinking the beginning of the Intro needs a bit more of a 
> feed-in.  Here's a (very) rough draft:

That's an idea.  As you no doubt saw, I began with "Testing bogofilter
(continued), because I was thinking in terms of linking to this from
the section on testing that's already in my bogofilter web page.  For
the version to be sent to the list, I could borrow a few words from
said web page, for example thus:

> The original version of bogofilter uses 
> the computation method presented in Paul Graham's paper (name & 
> url).  Gary Robinson took an interest in Graham's paper, and wrote
  an insightful commentary (url) in which he presented several suggestions
  (untested) for improvements to Graham's method.  I was intrigued,
  and modified bogofilter 0.7 (and subsequently 0.7.4 and 0.7.5) to try
  them out.  Initial tests (my page url) looked promising.  Discussion
  with David Smythe led to consensus that further testing would be
  worthwhile; we were interested in answering (more completely) the
  following two questions: yadda yadda yadda...

> Gotta go.  The neighborhood trick or treaters are starting to arrive.

It rained here, so only a few hardy little souls wandered soggily
round...

-- 
| G r e g  l a m b          | gpg public key:      |
|   http://www.sbc.nu/~glamb |   finger [e-mail] |

From [e-mail]  Fri Nov  1 06:32:06 2002
Return-Path: <[e-mail]>
Delivered-To: [e-mail]
Received: from oak.example.com (oak.example.com [219.168.1.10])
	by example.com (Postfix) with ESMTP id 9156627F82
	for <[e-mail]>; Fri,  1 Nov 2002 06:32:05 -0500 (EST)
Received: by oak.example.com (Postfix, from userid 1000)
	id 163A317E906; Fri,  1 Nov 2002 06:32:05 -0500 (EST)
To: [e-mail]
Subject: Hunger Site - Nov 01 - 1.1 cups
Message-Id: <[e-mail]>
Date: Fri,  1 Nov 2002 06:32:05 -0500 (EST)
From: [e-mail] (David Smythe)

1 this week; 1 in the last 7 days
1 this month; 1 in the last 30 days
1 this year; 1 since June 15th

From [e-mail]  Fri Nov  1 06:38:06 2002
Return-Path: <[e-mail]>
Delivered-To: [e-mail]
Received: from csl2.contronics.on.ca (csl2.contronics.on.ca [240.138.93.2])
	by example.com (Postfix) with ESMTP id B081C27F82
	for <[e-mail]>; Fri,  1 Nov 2002 06:38:05 -0500 (EST)
Received: from athame.dynagen.internal(localhost[127.0.0.1]) (1720 bytes) by csl2.contronics.on.ca
	via sendmail with P:esmtp/R:bind_hosts/T:inet_zone_bind_smtp
	(sender: <[e-mail]>) 
	id <[e-mail]>
	for <[e-mail]>; Fri, 1 Nov 2002 06:38:04 -0500 (EST)
	(Smail-3.2.0.114 2001-Aug-6 #5 built 2002-Mar-26)
Received: (1351 bytes) by athame.dynagen.internal
	via sendmail with P:stdio/R:smart_host/T:smtp
	(sender: <[e-mail]>) 
	id <[e-mail]>
	for [e-mail]; Fri, 1 Nov 2002 06:38:04 -0500 (EST)
	(Smail-3.2.0.111 2000-Feb-17 #5 built 2000-Apr-12)
Date: Fri, 1 Nov 2002 06:38:04 -0500
From: Greg Lamb <[e-mail]>
To: David Smythe <[e-mail]>
Subject: testing with training data
Message-ID: <[e-mail]>
Reply-To: Greg Lamb <[e-mail]>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Organization: Dynagen Consulting Limited

As you probably realized, my results in this latest round of testing,  
if they hold up, invalidate a lot of work that I and others have done.
It's very clear (and I will revise the discussion to say so explicitly) 
that the Robinson method, in particular, shouldn't be tested by having 
it re-evaluate its own training data.  If we do that, Robinson will
look good, but the test will have no value in predicting the
performance of the Robinson method in production, where most messages  
are new.
   
I'm not very surprised at this; I've always known that using training  
data for testing was a dangerous expedient.  It's impressive _how_
dangerous it turns out to be in the case of Robinson's approach,
though. 

Regards..............
-- 
| G r e g  l a m b          | gpg public key:      |
|   http://www.sbc.nu/~glamb |   finger [e-mail] |

From Online#[e-mail]  Fri Nov  1 07:08:32 2002
Return-Path: <Online#[e-mail]>
Delivered-To: [e-mail]
Received: from ABV-SFO1-ACMTA6.CNET.COM (abv-sfo1-acmta6.cnet.com [208.16.1.169])
	by example.com (Postfix) with ESMTP id AC46827F82
	for <[e-mail]>; Fri,  1 Nov 2002 07:08:30 -0500 (EST)
Received: from abv-sfo1-ac-agent6 (208.16.0.241) by ABV-SFO1-ACMTA6.CNET.COM (PowerMTA(TM) v1.5); Fri, 1 Nov 2002 04:12:31 -0800 (envelope-from <Online#[e-mail]>)
Message-ID: <1961754.1036152508913.JavaMail.root@abv-sfo1-ac-agent6>
Date: Fri, 1 Nov 2002 04:08:28 -0800 (PST)
From: ZDNet Tech Update Today <Online#[e-mail]>
To: [e-mail]
Subject: The hidden gotcha of spam [TECH UPDATE]
Mime-Version: 1.0
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
X-Mailer: Accucast (http://www.accucast.com)
X-Mailer-Version: 2.8.6-4

<html>
<head>
<title>Tech Update Today</title>
<link rel="stylesheet" type="text/css" href="http://home.cnet.com/css/all.css" />
<link rel="stylesheet" type="text/css" href="http://home.cnet.com/css/ie.css" />
<link rel="stylesheet" type="text/css" href="http://home.cnet.com/css/92.css" />
</head>
<body style="margin:8px 9px 9px 12px" bgcolor="#ffffff" background="http://techupdate.zdnet.com/techupdate/i/bg_232850.gif" link="#003399" alink="#cc0000" vlink="#666699">

<div align="center">

<!-- main -->