Rescore Mass-Check

This is the procedure we use to generate new scores. It takes quite a while and is labour-intensive, so we do it infrequently.

We generate new scores by analyzing a massive collection of mail (a "corpus"), and running software to create a score-set that gets the best possible set of scores, so that the maximum possible number of mails in that corpus are correctly classified (ie. so that SA thinks the ham messages are nearly all ham, and the spam messages are nearly all spam).

As MattKettler explains:

The corpus is a LOT (aprox 1million pieces) of real-world, hand sorted mail.

Basically a smallish number (less than 100) people, including the
developers themselves, work as volunteer "corpus submitters". They hand
sort their mail and run mass-check over it. They submit the output logs
mass-check generates. Occasionally people review the submitted logs for
obvious mistakes, but it is largely a trust system.

If you want to see the statistics from the last corpus run, check the
STATISTICS.txt files that come in the SA tarball. It will tell you how many
emails were used, and what the hit rates of all the rules were.

Procedure

Here's the process for generating the score-set.

TODO: this is no longer accurate – iirc we can do all mass-checks in one sitting. Daniel, can you update this?

1. heads-up

Inform everyone in advance on the -users and -dev lists that we will be starting mass-checks shortly, and they should get their corpora nice and clean.

2. announce mass-check run 1 (score sets 0 and 1)

See MassCheck. The mass-check for both scoresets can be done in one command, e.g.

  cd masses
  echo "use_bayes 0" > spamassassin/user_prefs
  mass-check --net [targets etc]

Here's the full announcement text for this phase: RescoreSet01Details

We then take the log files rsync'd up to the server, and use those logs for both set 0 and set 1; set 0 can be generated from set 1 by stripping out the network tests.

3. allow several days to complete

Provide enough time, including a weekend if possible, giving people enough time to get around to running it given that they may be busy with day-job stuff

4. generate scores for score sets 0 and 1

See RunningPerceptron.

Once this is complete, update rules/50_scores.cf with the generated scores.

5. announce mass-check run 2 (set 2) and run 3 (set 3)

See MassCheck. Because set 2 and set 3 both require scores from set 0 and set 1, and both depend on auto-learning, they have to be run separately.

Scoreset 2:

  cd masses
  echo "use_bayes 1" > spamassassin/user_prefs
  mass-check [targets etc]

Scoreset 3:

  cd masses
  echo "use_bayes 1" > spamassassin/user_prefs
  mass-check --net [targets etc]

Here's the full announcement text for this phase: RescoreSet23Details

6. wait for everyone to complete them, as per step #3.

Waiting...

7. generate scores for score sets 2 and 3