Hi all,
I use dspam-3.8.0, with postfix, myql-5.0.22, and I compiled dspam with
--enable-virtual-users. I set up a global group to provide
"out-of-the-box-filtering" for my users.
grupoJA:classification:[EMAIL PROTECTED]
I trained this global user with 1500 ham and 1500 spam, using dspam_train, with
this result:
Training Snapshot:
[EMAIL PROTECTED] TP: 1494 TN: 1533 FP: 6 FN: 82 SC: 25 NC: 2
SHR: 94.80% HSR: 0.39% OCA: 97.17%
Overall Statistics:
[EMAIL PROTECTED] TP: 1494 TN: 1533 FP: 6 FN: 82 SC: 25 NC: 2
SHR: 94.80% HSR: 0.39% OCA: 97.17%
This training result seemed fine.
Then I send a few ham mails to one of my user, all classified spam. This is one
mail's headers:
Subject: [SPAM] =?gb2312?B?suLK1Gh4t6LTyrz+IDA4MDUwOCAwMQ==?=
Date: Thu, 8 May 2008 10:24:23 +0800
MIME-Version: 1.0
Content-Type: multipart/alternative;
boundary="----=_NextPart_000_000E_01C8B0F5.AF9F7C00"
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 6.00.2800.1807
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1807
X-DSPAM-Result: Spam
X-DSPAM-Processed: Thu May 8 10:22:05 2008
X-DSPAM-Confidence: 0.6000
X-DSPAM-Probability: 0.0023
X-DSPAM-Factors: 27,
Received*hexun.com>, 0.40000,
equiv=Content+Type, 0.40000,
Date*May, 0.40000,
2800, 0.40000,
01+<META, 0.40000,
Received*ESMTP, 0.40000,
Subject*01, 0.40000,
Date*May+2008, 0.40000,
From*"dspam" <[EMAIL PROTECTED]>, 0.40000,
X-MimeOLE*By, 0.40000,
message, 0.40000,
Received*8+May, 0.40000,
Date*0800, 0.40000,
Received*(messenger, 0.40000,
From*"dspam", 0.40000,
X-MimeOLE*Produced, 0.40000,
Message-ID*hejunsong0680>, 0.40000,
080508, 0.40000,
080508, 0.40000,
<META+content="MSHTML, 0.40000,<BR> Received*(unknown+[202.99.16.20]),
0.40000,<BR> 080508+01, 0.40000,<BR> 080508+01, 0.40000,<BR> Received*Thu+8,
0.40000,<BR> multi, 0.40000,<BR> Subject*测试hx发邮件+080508, 0.40000,<BR>
From*"dspam"+<kirk020, 0.4</DIV>
<DIV> </DIV>
<DIV>should I just retrain these mails, or have I configured wrongly? Following
is the different options with standard dspam.conf:</DIV>
<DIV> </DIV>
<DIV>#TrustedDeliveryAgent "/usr/bin/procmail"<BR>DeliveryHost
127.0.0.1<BR>DeliveryPort 10027<BR>DeliveryIdent
localhost<BR>DeliveryProto SMTP<BR>Trust postfix<BR>Debug
*<BR>TrainingMode teft<BR>Preference "signatureLocation=headers"<BR>Preference
"spamAction=tag"<BR>MySQLServer
/tmp/mysql.sock<BR>MySQLPort<BR>MySQLUser dspam<BR>MySQLPass
changeme<BR>MySQLDb dspam<BR>MySQLCompress true<BR>ServerPID
/var/run/dspam.pid<BR>ServerMode auto<BR>ServerParameters
"--deliver=innocent"<BR>ServerIdent
"localhost.localdomain"<BR>ServerDomainSocketPath "/tmp/dspam.sock"<BR></DIV>
<DIV>I also attached debug file dspam.debug. </DIV>
<DIV>What should i do to gain good accuracy?</DIV>
<DIV> </DIV>
<DIV>Thanks!</DIV>
<DIV> kirk</DIV>
<DIV> </DIV>
<DIV> </DIV>
送！送！送！正版瑞星2008半年免费！
!DSPAM:1011,482281d6122631643014811!
903: [05/08/2008 10:22:05] DSPAM Instance Startup
903: [05/08/2008 10:22:05] input args: dspam --deliver=innocent
903: [05/08/2008 10:22:05] pass-thru args:
903: [05/08/2008 10:22:05] processing user [EMAIL PROTECTED]
903: [05/08/2008 10:22:05] uid = 0, euid = 0, gid = 0, egid = 12
903: [05/08/2008 10:22:05] loading preferences for user [EMAIL PROTECTED]
903: [05/08/2008 10:22:05] Loading preferences for uid 2
903: [05/08/2008 10:22:05] Loading preferences for uid 0
903: [05/08/2008 10:22:05] Loading preferences for uid 0
903: [05/08/2008 10:22:05] default preferences empty. reverting to dspam.conf
preferences.
903: [05/08/2008 10:22:05] Loading preferences from dspam.conf
903: [05/08/2008 10:22:05] using /opt/dspam/var/dspam/opt-in/[EMAIL PROTECTED]
as path
903: [05/08/2008 10:22:05] using /opt/dspam/var/dspam/opt-out/[EMAIL PROTECTED]
as path
903: [05/08/2008 10:22:05] adding user [EMAIL PROTECTED] to classification
group grupoJA
903: [05/08/2008 10:22:05] sedation level set to: 0
903: [05/08/2008 10:22:05] message is signed. retaining original text for
reassembly
903: [05/08/2008 10:22:05] message is signed. retaining original text for
reassembly
903: [05/08/2008 10:22:05] Whitelist threshold: 10
903: [05/08/2008 10:22:05] [graham] [0.400000] Received*hexun.com> (1frq, 4s,
0i)
903: [05/08/2008 10:22:05] [burton] [0.400000] Received*hexun.com> (1frq, 4s,
0i)
903: [05/08/2008 10:22:05] [graham] [0.400000] equiv=Content+Type (1frq, 4s, 0i)
903: [05/08/2008 10:22:05] [burton] [0.400000] equiv=Content+Type (1frq, 4s, 0i)
903: [05/08/2008 10:22:05] [graham] [0.400000] Date*May (1frq, 4s, 0i)
903: [05/08/2008 10:22:05] [burton] [0.400000] Date*May (1frq, 4s, 0i)
903: [05/08/2008 10:22:05] [graham] [0.400000] 2800 (1frq, 4s, 0i)
903: [05/08/2008 10:22:05] [burton] [0.400000] 2800 (1frq, 4s, 0i)
903: [05/08/2008 10:22:05] [graham] [0.400000] 01+<META (1frq, 0s, 0i)
903: [05/08/2008 10:22:05] [burton] [0.400000] 01+<META (1frq, 0s, 0i)
903: [05/08/2008 10:22:05] [graham] [0.400000] Received*ESMTP (1frq, 4s, 0i)
903: [05/08/2008 10:22:05] [burton] [0.400000] Received*ESMTP (1frq, 4s, 0i)
903: [05/08/2008 10:22:05] [graham] [0.400000] Subject*01 (1frq, 0s, 0i)
903: [05/08/2008 10:22:05] [burton] [0.400000] Subject*01 (1frq, 0s, 0i)
903: [05/08/2008 10:22:05] [graham] [0.400000] Date*May+2008 (1frq, 4s, 0i)
903: [05/08/2008 10:22:05] [burton] [0.400000] Date*May+2008 (1frq, 4s, 0i)
903: [05/08/2008 10:22:05] [graham] [0.400000] From*"dspam" <[EMAIL PROTECTED]>
(1frq, 2s, 0i)
903: [05/08/2008 10:22:05] [burton] [0.400000] From*"dspam" <[EMAIL PROTECTED]>
(1frq, 2s, 0i)
903: [05/08/2008 10:22:05] [graham] [0.400000] X-MimeOLE*By (1frq, 2s, 0i)
903: [05/08/2008 10:22:05] [burton] [0.400000] X-MimeOLE*By (1frq, 2s, 0i)
903: [05/08/2008 10:22:05] [graham] [0.400000] message (1frq, 4s, 0i)
903: [05/08/2008 10:22:05] [burton] [0.400000] message (1frq, 4s, 0i)
903: [05/08/2008 10:22:05] [graham] [0.400000] Received*8+May (1frq, 0s, 0i)
903: [05/08/2008 10:22:05] [burton] [0.400000] Received*8+May (1frq, 0s, 0i)
903: [05/08/2008 10:22:05] [graham] [0.400000] Date*0800 (1frq, 4s, 0i)
903: [05/08/2008 10:22:05] [burton] [0.400000] Date*0800 (1frq, 4s, 0i)
903: [05/08/2008 10:22:05] [graham] [0.400000] Received*(messenger (1frq, 4s,
0i)
903: [05/08/2008 10:22:05] [burton] [0.400000] Received*(messenger (1frq, 4s,
0i)
903: [05/08/2008 10:22:05] [graham] [0.400000] From*"dspam" (1frq, 2s, 0i)
903: [05/08/2008 10:22:05] [burton] [0.400000] From*"dspam" (1frq, 2s, 0i)
903: [05/08/2008 10:22:05] [burton] [0.400000] X-MimeOLE*Produced (1frq, 2s, 0i)
903: [05/08/2008 10:22:05] [burton] [0.400000] Message-ID*hejunsong0680> (1frq,
2s, 0i)
903: [05/08/2008 10:22:05] [burton] [0.400000] 080508 (2frq, 0s, 0i)
903: [05/08/2008 10:22:05] [burton] [0.400000] 080508 (2frq, 0s, 0i)
903: [05/08/2008 10:22:05] [burton] [0.400000] <META+content="MSHTML (1frq, 4s,
0i)
903: [05/08/2008 10:22:05] [burton] [0.400000]
Received*(unknown+[202.99.16.20]) (1frq, 4s, 0i)
903: [05/08/2008 10:22:05] [burton] [0.400000] 080508+01 (2frq, 0s, 0i)
903: [05/08/2008 10:22:05] [burton] [0.400000] 080508+01 (2frq, 0s, 0i)
903: [05/08/2008 10:22:05] [burton] [0.400000] Received*Thu+8 (1frq, 0s, 0i)
903: [05/08/2008 10:22:05] [burton] [0.400000] multi (1frq, 4s, 0i)
903: [05/08/2008 10:22:05] [burton] [0.400000] Subject*²âÊÔhx·¢Óʼþ+080508
(1frq, 0s, 0i)
903: [05/08/2008 10:22:05] [burton] [0.400000] From*"dspam"+<kirk020 (1frq, 2s,
0i)
903: [05/08/2008 10:22:05] Graham-Bayesian Probability: 0.002278 Samples: 15
903: [05/08/2008 10:22:05] Burton-Bayesian Probability: 0.000018 Samples: 27
903: [05/08/2008 10:22:05] no factors specified; using default
903: [05/08/2008 10:22:05] Result Confidence: 1.00
903: [05/08/2008 10:22:05] Control: [10 10] [10 11] Delta: [0 1]
903: [05/08/2008 10:22:05] total processing time: 0.02881s
903: [05/08/2008 10:22:05] checking result for user [EMAIL PROTECTED]
903: [05/08/2008 10:22:05] [graham] [0.994013] E: 5483309457905308941 (1frq,
345s, 1i)
903: [05/08/2008 10:22:05] [burton] [0.994013] E: 5483309457905308941 (1frq,
345s, 1i)
903: [05/08/2008 10:22:05] [graham] [0.010000] E: 4293821448246082252 (1frq,
0s, 7i)
903: [05/08/2008 10:22:05] [burton] [0.010000] E: 4293821448246082252 (1frq,
0s, 7i)
903: [05/08/2008 10:22:05] [graham] [0.990000] E: 12016759050104786585 (1frq,
116s, 0i)
903: [05/08/2008 10:22:05] [burton] [0.990000] E: 12016759050104786585 (1frq,
116s, 0i)
903: [05/08/2008 10:22:05] [graham] [0.990000] E: 3255294038965747712 (1frq,
110s, 0i)
903: [05/08/2008 10:22:05] [burton] [0.990000] E: 3255294038965747712 (1frq,
110s, 0i)
903: [05/08/2008 10:22:05] [graham] [0.990000] E: 9911768209844530249 (1frq,
116s, 0i)
903: [05/08/2008 10:22:05] [burton] [0.990000] E: 9911768209844530249 (1frq,
116s, 0i)
903: [05/08/2008 10:22:05] [graham] [0.990000] E: 6867228121829851563 (1frq,
116s, 0i)
903: [05/08/2008 10:22:05] [burton] [0.990000] E: 6867228121829851563 (1frq,
116s, 0i)
903: [05/08/2008 10:22:05] [graham] [0.010000] E: 9273557869598762556 (1frq,
0s, 9i)
903: [05/08/2008 10:22:05] [burton] [0.010000] E: 9273557869598762556 (1frq,
0s, 9i)
903: [05/08/2008 10:22:05] [graham] [0.010000] E: 14875237204334870012 (1frq,
0s, 5i)
903: [05/08/2008 10:22:05] [burton] [0.010000] E: 14875237204334870012 (1frq,
0s, 5i)
903: [05/08/2008 10:22:05] [graham] [0.990000] E: 3255321295901949952 (1frq,
131s, 0i)
903: [05/08/2008 10:22:05] [burton] [0.990000] E: 3255321295901949952 (1frq,
131s, 0i)
903: [05/08/2008 10:22:05] [graham] [0.990000] E: 17877205250337029371 (1frq,
116s, 0i)
903: [05/08/2008 10:22:05] [burton] [0.990000] E: 17877205250337029371 (1frq,
116s, 0i)
903: [05/08/2008 10:22:05] [graham] [0.010000] E: 13348477354612403345 (1frq,
0s, 3i)
903: [05/08/2008 10:22:05] [burton] [0.010000] E: 13348477354612403345 (1frq,
0s, 3i)
903: [05/08/2008 10:22:05] [graham] [0.010000] E: 9273557857864649442 (1frq,
0s, 7i)
903: [05/08/2008 10:22:05] [burton] [0.010000] E: 9273557857864649442 (1frq,
0s, 7i)
903: [05/08/2008 10:22:05] [graham] [0.010000] E: 13573258275136642356 (1frq,
0s, 11i)
903: [05/08/2008 10:22:05] [burton] [0.010000] E: 13573258275136642356 (1frq,
0s, 11i)
903: [05/08/2008 10:22:05] [graham] [0.010000] E: 9206198621879547341 (1frq,
0s, 3i)
903: [05/08/2008 10:22:05] [burton] [0.010000] E: 9206198621879547341 (1frq,
0s, 3i)
903: [05/08/2008 10:22:05] [graham] [0.990000] E: 3255294038856510720 (1frq,
107s, 0i)
903: [05/08/2008 10:22:05] [burton] [0.990000] E: 3255294038856510720 (1frq,
107s, 0i)
903: [05/08/2008 10:22:05] [burton] [0.010000] E: 11663502761777012730 (1frq,
0s, 5i)
903: [05/08/2008 10:22:05] [burton] [0.990000] E: 14811854249704039829 (1frq,
126s, 0i)
903: [05/08/2008 10:22:05] [burton] [0.010000] E: 9365566577174198171 (1frq,
0s, 10i)
903: [05/08/2008 10:22:05] [burton] [0.990000] E: 10878069382963375208 (1frq,
126s, 0i)
903: [05/08/2008 10:22:05] [burton] [0.984148] E: 7908790535211450368 (2frq,
129s, 1i)
903: [05/08/2008 10:22:05] [burton] [0.984148] E: 7908790535211450368 (2frq,
129s, 1i)
903: [05/08/2008 10:22:05] [burton] [0.982403] E: 13028519796621377271 (1frq,
116s, 1i)
903: [05/08/2008 10:22:05] [burton] [0.982403] E: 803101134672876591 (1frq,
116s, 1i)
903: [05/08/2008 10:22:05] [burton] [0.980035] E: 5049287340227680834 (1frq,
102s, 1i)
903: [05/08/2008 10:22:05] [burton] [0.975505] E: 6230608064996248496 (1frq,
331s, 4i)
903: [05/08/2008 10:22:05] [burton] [0.975360] E: 4761609862965000081 (1frq,
329s, 4i)
903: [05/08/2008 10:22:05] [burton] [0.975360] E: 4820269125774155977 (1frq,
329s, 4i)
903: [05/08/2008 10:22:05] Graham-Bayesian Probability: 0.994013 Samples: 15
903: [05/08/2008 10:22:05] Burton-Bayesian Probability: 1.000000 Samples: 27
903: [05/08/2008 10:22:05] using Graham factors
903: [05/08/2008 10:22:05] Result Confidence: 0.54
903: [05/08/2008 10:22:05] total processing time: 0.01253s
903: [05/08/2008 10:22:05] CLASSIFY CATCH: [EMAIL PROTECTED]
903: [05/08/2008 10:22:05] processing signature. length: 1764
903: [05/08/2008 10:22:05] reversing 147 tokens
903: [05/08/2008 10:22:05] Control: [10 10] [11 9] Delta: [1 -1]
903: [05/08/2008 10:22:05] saving signature as 482263cd9031681412463
903: [05/08/2008 10:22:05] libdspam returned probability of 0.002278
903: [05/08/2008 10:22:05] message result: SPAM
903: [05/08/2008 10:22:05] Establishing connection to 127.0.0.1:10027
903: [05/08/2008 10:22:05] Connection established
903: [05/08/2008 10:22:06] DSPAM Instance Shutdown. Exit Code: 0
903: [05/08/2008 10:22:06] checking trusted user list for root(0)
