Corpora: Re: multiple-category classification of text

J. Zavrel Jakub.Zavrel at kub.nl
Thu May 11 23:00:22 UTC 2000


On Thu, 11 May 2000, Seth Russell wrote:
> yannick at planetarymotion.net wrote:
>
> > I'm looking for references on multiple-category classification of
> > text.

Hi,

Just couldn't resist putting in a small shameless plug for my
favorite machine learning flavor:

This type of text classification can be done very easily using
memory-based learning. If the nearest neighbors of a phrase to be
classified are multiple-category, you will also be able to assign
multiple-categories (i.e. a distribution of them). By controlling the
number of nearest neighbors you can systematically increase the number of
suggested categories. But of course this can be done using Naive Bayes as
well...

It's easy to try with our memory-based learning package TiMBL
which can be obtained from http://ilk.kub.nl (follow the link to
software).

Cheers,

--Jakub

------------------------------------------------------------------------------
Jakub Zavrel, B 330, Tilburg University, POBox 90153, 5000 LE Tilburg, NL
http://ilk.kub.nl/~zavrel/  tel/fax: +31-13-4663163/3110
------------------------------------------------------------------------------



More information about the Corpora mailing list