[Spambayes] CL3 results vs. CL2

[Brad Clements]

Didn't change anything from the CL2 test. I haven't had a chance to examine the new 
false negatives.

CL3 Results:

/tmp/clgen-cl3-5x5 -> /tmp/clgen-cl3-5x5s.txt
-> Training on Data/Ham/Set{6,7,8,9,10} & Data/Spam/Set{6,7,8,9,10} ... 6500 hams & 6500 spams
-> <stat> population hammean -0.353152965041 hamvar 0.0190809424597
-> <stat> population spammean -0.230738145294 spamvar 0.0141618505748
-> Predicting Data/Ham/Set{1,2,3,4,5} & Data/Spam/Set{1,2,3,4,5} ...
-> <stat> tested 6500 hams & 6500 spams against 6500 hams & 6500 spams
-> <stat> false positive %: 0.8
-> <stat> false negative %: 0.569230769231
      0.800   0.569
-> <stat> 52 new false positives
-> <stat> 37 new false negatives
-> <stat> Ham scores for all in this training set: 6500 items; mean 1.11; sdev 8.23
-> <stat> min 0; median 0; max 100
-> <stat> Spam scores for all in this training set: 6500 items; mean 98.96; sdev 7.46
-> <stat> min 0; median 100; max 100
-> best cutoff for all in this training set: 0.5
->     with weighted total 1*52 fp + 37 fn = 89
->     fp rate 0.8%  fn rate 0.569%
-> <stat> Ham scores for all runs: 6500 items; mean 1.11; sdev 8.23
-> <stat> min 0; median 0; max 100
-> <stat> Spam scores for all runs: 6500 items; mean 98.96; sdev 7.46
-> <stat> min 0; median 100; max 100
-> best cutoff for all runs: 0.5
->     with weighted total 1*52 fp + 37 fn = 89
->     fp rate 0.8%  fn rate 0.569%
total unique false pos 52
total unique false neg 37
average fp % 0.8
average fn % 0.569230769231


CL2 Results:

/tmp/clgen-cl2-5x5 -> /tmp/clgen-cl2-5x5s.txt
-> Training on Data/Ham/Set{6,7,8,9,10} & Data/Spam/Set{6,7,8,9,10} ... 6500 hams & 6500 spams
-> <stat> population hammean -0.258919766598 hamvar 0.235232283813
-> <stat> population spammean -0.238803626095 spamvar 0.189273495163
-> Predicting Data/Ham/Set{1,2,3,4,5} & Data/Spam/Set{1,2,3,4,5} ...
-> <stat> tested 6500 hams & 6500 spams against 6500 hams & 6500 spams
-> <stat> false positive %: 1.12307692308
-> <stat> false negative %: 0.369230769231
      1.123   0.369
-> <stat> 73 new false positives
-> <stat> 24 new false negatives
-> <stat> Ham scores for all in this training set: 6500 items; mean 1.53; sdev 9.48
-> <stat> min 0; median 0; max 100
-> <stat> Spam scores for all in this training set: 6500 items; mean 99.17; sdev 6.93
-> <stat> min 0; median 100; max 100
-> best cutoff for all in this training set: 0.5
->     with weighted total 1*73 fp + 24 fn = 97
->     fp rate 1.12%  fn rate 0.369%
-> <stat> Ham scores for all runs: 6500 items; mean 1.53; sdev 9.48
-> <stat> min 0; median 0; max 100
-> <stat> Spam scores for all runs: 6500 items; mean 99.17; sdev 6.93
-> <stat> min 0; median 100; max 100
-> best cutoff for all runs: 0.5
->     with weighted total 1*73 fp + 24 fn = 97
->     fp rate 1.12%  fn rate 0.369%
total unique false pos 73
total unique false neg 24
average fp % 1.12307692308
average fn % 0.369230769231



Brad Clements,                bkc@murkworks.com   (315)268-1000
http://www.murkworks.com                          (315)268-9812 Fax
AOL-IM: BKClements