what was used for this:
OVERALL% SPAM% HAM% S/O RANK SCORE NAME
66968 41748 25220 0.623 0.00 0.00 (all messages)
882 882 0 1.000 1.00 0.75 SARE_OEM_PRODS_1
700 700 0 1.000 0.94 1.00 SARE_OEM_OBFU
606 606 0 1.000 0.89 2.00 SARE_OEM_A_1
1586 1585 1 0.999 0.84 0.88 SARE_PRODUCTS_03
523 523 0 1.000 0.83 1.90 SARE_PRODS_LOTS
523 523 0 1.000 0.83 1.75 SARE_PRODUCTS_04
498 498 0 1.000 0.77 1.50 SARE_OEM_A_2
404 404 0 1.000 0.72 0.90 SARE_OEM_PRODS_2
402 402 0 1.000 0.66 1.00 SARE_OEM_SOFT_IS
326 326 0 1.000 0.60 0.95 SARE_OEM_PRODS_3
260 260 0 1.000 0.55 1.70 SARE_OEM_FAKE_YEAR
249 249 0 1.000 0.49 1.26 SARE_OEM_AND_OTHER
2128 2121 7 0.995 0.47 0.88 SARE_OEM_PRODS_FEW
59 59 0 1.000 0.37 1.00 SARE_OEM_S_PRICE
27 27 0 1.000 0.32 2.00 SARE_OEM_B_3
2133 2125 8 0.994 0.31 0.38 SARE_PRODUCTS_02
13 13 0 1.000 0.26 0.80 SARE_OEM_OEMCD
4 4 0 1.000 0.20 0.80 SARE_OEM_REDPR
0 0 0 0.500 0.15 1.00 SARE_OEM_BRC
78 74 4 0.918 0.00 1.20 SARE_OEM_S_DOL
On Apr 20, 2006, at 3:37 PM, Benjamin Adams wrote:
I'm going to start changing rules but would like to get away to
view stats first.
Something like:
Past 7 day stats for USERNAME:
RULES Mark SPAM Mark HAM
HTML_MESSAGE 28 20
UNPARSEABLE_RELAY 31 50
MIME_BASE64_TEXT 34 23
That way I can say 'I have to sa-learn 17 messages that have
HTML_MASSAGE tag but not marked SPAM'
I can change the score some on the HTML_MESSAGE and more will be
marked as SPAM.
Viewing the big picture rather then going through messages one by
one and making my own stats.
Is this a bad idea?
Ben
