SA Bugzilla – Bug 5284
MIME_BOUND_NEXTPART has bad performance
Last modified: 2007-01-09 23:26:13 UTC
In the 3.1 score gen mass-check run, this rule didn't do terribly, but also wasn't great. At this point, the rule is basically useless (from last night's mass-check run): 0.452 0.4522 0.4530 0.500 0.50 0.00 MIME_BOUND_NEXTPART So we either need to do some more work and fix this, or just drop the rule.
wow, how did that not get removed during the big rule pruning? I vote for dropping it.
we don't need to resolve this before mass-checks, actually; if needs be, we can remove the rule just before running the perceptron (and just ignore the hits in the logs). updating status accordingly
ok, the performance went down since the last time I checked: 27.865 31.7711 3.2772 0.906 0.59 0.00 __NEXTPART_NORMAL 28.172 32.0449 3.7978 0.894 0.58 0.00 __NEXTPART_ALL 0.308 0.2737 0.5206 0.345 0.45 0.00 MIME_BOUND_NEXTPART so yeah, it's gone.