tuning.sh - saner values

David Relson relson at osagesoftware.com
Sat Jun 14 20:23:01 CEST 2003


Hi,

Given all the traffic between Greg and pi, I'm making some corrections and 
changes to the script.  As corrections, I'm fixing the calls to getco and 
wrapper and exporting "target" so that perl can use it.  I've deleted "1" 
from the robs values tested because it's simply too big and 0.475 from 
min_dev values tested because a value that large can exclude so many tokens 
from the spamicity calculation that the final score may be based on only 2 
or 3 tokens - which is too few for safety.

Here're some results that show the need for deleting those values.

           r0     r1     r2
sp.mc    179    179    179
ns.mc    845    845    845

bogofilter version 0.13.6.2.cvs.20030606
robx        = 0.415000 (4.15e-01)
robs        = 1.000000 (1.00e+00)
min_dev     = 0.100000 (1.00e-01)
06/14 14:02:24   1       0.025 fpos...7 at cutoff 0.870000, run0... 
58  run1... 62  run2... 66  186
06/14 14:02:29   1       0.050 fpos...7 at cutoff 0.882000, run0... 
56  run1... 60  run2... 67  183
06/14 14:02:34   1       0.075 fpos...7 at cutoff 0.908400, run0... 
53  run1... 55  run2... 63  171
06/14 14:02:40   1       0.100 fpos...7 at cutoff 0.945700, run0... 
65  run1... 63  run2... 72  200
06/14 14:02:45   1       0.125 fpos...7 at cutoff 0.956000, run0... 
63  run1... 58  run2... 69  190
06/14 14:02:50   1       0.150 fpos...7 at cutoff 0.959400, run0... 
63  run1... 61  run2... 70  194
06/14 14:02:55   1       0.175 fpos...7 at cutoff 0.960900, run0... 
59  run1... 58  run2... 65  182
06/14 14:03:01   1       0.200 fpos...7 at cutoff 0.960800, run0... 
58  run1... 55  run2... 62  175
06/14 14:03:06   1       0.225 fpos...7 at cutoff 0.959300, run0... 
60  run1... 56  run2... 64  180
06/14 14:03:10   1       0.250 fpos...7 at cutoff 0.959300, run0... 
60  run1... 57  run2... 64  181
06/14 14:03:16   1       0.275 fpos...7 at cutoff 0.965600, run0... 
67  run1... 63  run2... 69  199
06/14 14:03:21   1       0.300 fpos...7 at cutoff 0.956900, run0... 
64  run1... 63  run2... 69  196
06/14 14:03:26   1       0.325 fpos...7 at cutoff 0.902600, run0... 
72  run1... 78  run2... 84  234
06/14 14:03:31   1       0.350 fpos...7 at cutoff 0.827000, run0... 
72  run1... 74  run2... 79  225
06/14 14:03:36   1       0.375 fpos...7 at cutoff 0.974500, run0... 
97  run1... 90  run2...101  288
06/14 14:03:41   1       0.400 fpos...7 at cutoff 0.671000, run0... 
82  run1... 84  run2... 91  257
06/14 14:03:46   1       0.425 fpos...7 at cutoff 0.415000, 
run0...  3  run1...  8  run2...  9   20
06/14 14:03:51   1       0.450 fpos...7 at cutoff 0.415000, 
run0...  3  run1...  3  run2...  4   10
06/14 14:03:56   1       0.475 fpos...7 at cutoff 0.415000, 
run0...  1  run1...  3  run2...  2    6





More information about the Bogofilter mailing list