histogram [was: Accuracy is lacking]
David Relson
relson at osagesoftware.com
Fri Feb 14 14:58:41 CET 2003
Greetings,
FWIW, here's the score and the histogram for my previous message - the one
that included the scores and subjects for my January high and low scoring
"Unsures". Many of the 87 low scorers in the histogram are from the
mailing list's header and trailer lines. The low spamicity tokens were
_almost_ enough to put it below the ham_cutoff of 0.10 - but not quite.
X-Bogosity: Unsure, tests=bogofilter-f, spamicity=0.129247,
version=0.10.2.tst.20030213
Status:
# int cnt prob spamicity histogram
# 0.00 87 0.018467 0.006319
##################################################
# 0.10 30 0.151211 0.023873 ##################
# 0.20 29 0.258583 0.052192 #################
# 0.30 39 0.341578 0.098603 #######################
# 0.40 0 0.000000 0.098603
# 0.50 0 0.000000 0.098603
# 0.60 38 0.650043 0.187839 ######################
# 0.70 39 0.750240 0.271159 #######################
# 0.80 19 0.833209 0.309529 ###########
# 0.90 18 0.984386 0.400346 ###########
More information about the Bogofilter
mailing list