[Snowball-discuss] misleading directive in definition of Porter2 stemmer
Martin Porter
martin at porterloo.wanadoo.co.uk
Mon Sep 8 16:00:01 BST 2008
Sam,
Thanks fot that suggestion, and I've improved the wording (actually to
"delete if the preceding word part contains a vowel, and after the deletion:")
Martin
At 17:47 04/09/2008 -0400, Sam Pilato wrote:
>Hi, Martin,
>
>In your definition of the English (Porter2) stemming algorithm
(http://snowball.tartarus.org/algorithms/english/stemmer.html), in
>Step 1b, for the second set of suffixes, you have:
>
> "delete if the preceding word part contains a vowel, and then [...]"
>
>I think you mean to test the disjuncts that follow only if the deletion
succeeded. (Otherwise, the stemmer with the last disjunct
>stems "beds", "sheds", and "shreds" to "bede", "shede", and "shrede".) That
could be made more explicit by changing the wording to:
>
> "if the preceding word part contains a vowel, delete the suffix and then
[...]"
>
>Regards,
>Sam
More information about the Snowball-discuss
mailing list