hi
dpam_train question
How do we train dspam in advance using a corpus of good emails and spam emails -- using dspam_train
Since a balanced corpus training of spam and ham is required. we can have two seperate maildirs say : spam@mydomain.com and ham@mydomain.com. The emails will be those already processed by dspam and containing dspam signature and seperated out by us manually.
question 1)
Is the above correct ? Or do we need to train using pristine ham and spam messages prior to the installation of dspam ?
question 2)
As per the dspam man page dspam_train tool "will present each message to dspam for a classification and then retrain only if the ** message was incorrect **."
thanks
rajesh mahadevan
Received on Mon Mar 19 17:15:12 2007
This archive was generated by hypermail 2.1.8 : Wed Mar 21 2007 - 00:00:07 CET