[Snowball-discuss] two results
Mon Oct 6 13:30:01 2003
Continuing Richard's discussion, an example in English would be "routing",
which might stem either to "route" (a verb meaning "to direct"), or "rout"
(a verb meaning "to defeat in battle"). This example is given elsewhere on
the Snowball website. In English, examples like this are very rare.
Richard is right in saying the stemmers on the Snowball site reduce all
words to just one form, but other approaches are possible. For example the
Willett/Schinke Latin stemmer reduces each word to two forms, in the hope
that at least one of the forms is correct, and I have implemented it in
So it is certainly possible, although it doesn't fit the design scheme of
the stemmers currently distributed on the Snowball site.
P.S. Do you know about
M. Popovic and P. Willett. The Effectiveness of Stemming for Natural
Language Access to Slovene Textual Data. Journal of the American Society for
Information Science, 43(5):384-390, 1992
(I have not myself read this paper, as far as I can recall.)