histogram, a.k.a. verbosity level
David Relson
relson at osagesoftware.com
Sun Oct 27 17:24:56 CET 2002
Greetings,
As many of you already know, I've got reasonable printout levels for the
Graham algorithm for verbosity=1,2, etc. I've been working on comparable
functionality for the Robinson algorithm.
The attached file shows how bogofilter responds to the following sets of
options:
-g -v
-r -v
-g -vv
-r -vv
-r -vvv
The first two just print out the "X-Bogosity" line.
"-g -vv" prints out the Graham extrema data (15 most extreme tokens with
probability, cumulative spamicity, and token).
"-r -vv" prints out a Robinson histogram (showing interval, token count for
interval, avg probability for interval, cumulative spamicity, and hash mark
histogram).
"-r -vvv" prints out the Rtable. Alternatively, "-R bogo.tbl" could be
used (for a similar effect).
The outputs are all to stdout so that they can be included in the delivered
email. My personal copy of bogofilter runs with verbose=2, i.e. "-vv",
and some extra threshold values so messages with no spam indicators only
show the "X-Bogosity" line.
I'll probably be adding the code to cvs in a day or two.
David
-------------- next part --------------
**** bogofilter -g -v msg.test
X-Bogosity: No, tests=bogofilter, spamicity=0.002278, version=0.7.6-1027.1028
**** bogofilter -r -v msg.test
X-Bogosity: No, tests=bogofilter, spamicity=0.336295, version=0.7.6-1027.1028
pri**** bogofilter -g -vv msg.test
X-Bogosity: No, tests=bogofilter, spamicity=0.002278, version=0.7.6-1027.1028
# 0.400000 0.400000 63.121.31.110
# 0.400000 0.307692 delivered-to
# 0.400000 0.228571 flipdog.com
# 0.400000 0.164948 from
# 0.400000 0.116364 fs1.i.flipdog.com
# 0.400000 0.080706 ftp.flipdog.com
# 0.400000 0.055292 jobhunter
# 0.400000 0.037553 oct
# 0.400000 0.025353 osagesoftware.com
# 0.400000 0.017046 postfix
# 0.400000 0.011429 received
# 0.400000 0.007648 relson
# 0.400000 0.005112 return-path
# 0.400000 0.003414 sun
# 0.400000 0.002278 with
**** bogofilter -r -vv msg.test
X-Bogosity: No, tests=bogofilter, spamicity=0.336295, version=0.7.6-1027.1028
# int cnt prob spamicity histogram
# 0.00 40 0.041188 0.011873 ############################
# 0.10 35 0.144202 0.039459 ########################
# 0.20 42 0.250499 0.088020 #############################
# 0.30 39 0.347486 0.141013 ###########################
# 0.40 74 0.418468 0.231323 ##################################################
# 0.50 22 0.559556 0.262952 ###############
# 0.60 11 0.648215 0.281128 ########
# 0.70 13 0.761747 0.307962 #########
# 0.80 10 0.865302 0.333357 #######
# 0.90 1 0.919971 0.336295 #
**** bogofilter -r -vvv msg.test
X-Bogosity: No, tests=bogofilter, spamicity=0.336295, version=0.7.6-1027.1028
# Token pgood pbad fw invfwlog fwlog
# 0 detroit 171.08 0 0.000001 -0.00000 -13.65886
# 1 supporting 30.86 0 0.000007 -0.00001 -11.91842
# 2 gathering 10.16 0 0.000020 -0.00002 -10.81988
# 3 modular 8.59 0 0.000025 -0.00002 -10.59676
# 4 hills 7.94 0 0.000029 -0.00003 -10.46325
# 5 documenting 5.34 0 0.000040 -0.00004 -10.12683
...[snip]...
# 284 ffffff 280.71 2152 0.884868 -2.16168 -0.12232
# 285 helvetica 227.98 2017 0.898841 -2.29106 -0.10665
# 286 efefef 2.34 23 0.919971 -2.52537 -0.08341
# 287 P_Q_S_invsum_logsum 0.40138 0.79216 0.336295 -147.270 -450.877
-------------- next part --------------
--------------------------------------------------------
David Relson Osage Software Systems, Inc.
relson at osagesoftware.com Ann Arbor, MI 48103
www.osagesoftware.com tel: 734.821.8800
More information about the bogofilter-dev
mailing list