[Spambayes-checkins] spambayes timtoken.py,1.2,1.3
Tim Peters
tim_one@users.sourceforge.net
Fri, 06 Sep 2002 13:42:42 -0700
Update of /cvsroot/spambayes/spambayes
In directory usw-pr-cvs1:/tmp/cvs-serv11651
Modified Files:
timtoken.py
Log Message:
Added a note about an experiment with no lower limit on the length of
words we'll look at. Didn't matter to f-p, but hurt f-n.
Index: timtoken.py
===================================================================
RCS file: /cvsroot/spambayes/spambayes/timtoken.py,v
retrieving revision 1.2
retrieving revision 1.3
diff -C2 -d -r1.2 -r1.3
*** timtoken.py 6 Sep 2002 19:12:59 -0000 1.2
--- timtoken.py 6 Sep 2002 20:42:40 -0000 1.3
***************
*** 392,395 ****
--- 392,397 ----
# XXX runs -- overall, no significant difference. It's only "common
# XXX sense" so far driving the exclusion of lengths 1 and 2.
+ # XXX Later: A test with no lower bound showed a significant increase
+ # XXX in the f-n rate. Curious!
# Make sure this range matches in tokenize().