Meyer, Tony wrote:
>>Having dot spambayes working great for about a
>>week, it suddenly stop classifying email and I
>>got this exception in my email header
>>
>><quote>
>>X-Spambayes-Exception: Traceback (most recent call last):
> 
> [...]
> 
>>    line 308, in probability .    assert hamcount <= nham
>>   .AssertionError
> 
> 
> This means that the database is corrupt (there's a token that has been seen 
> in more ham than the total number of ham you have trained, which is 
> impossible).  The best thing to do is retrain from scratch.
> 
> It's pretty rare for this to occur these days.  How have you been training 
> SpamBayes?
> 
> =Tony.Meyer

Thanks very much, Tony. I now realize that.

My spam and ham mails are moved to separate mboxes and trained using
sb_mboxtrain.py. The mails are allowed to accumulated over time and I
did the training at the end of the day's work. I recalled I got over a
thousand of each and hammie.db grew to about 11 Mb. Perhaps I
overtrained. -(

Regards,
ST
--
_______________________________________________
[email protected]
http://mail.python.org/mailman/listinfo/spambayes
Check the FAQ before asking: http://spambayes.sf.net/faq.html

Reply via email to