Re: Corpora: Automatic recognition of product names

Bruce L. Lambert (lambertb@uic.edu)
Mon, 01 Nov 1999 10:35:47 -0600

One strategy would be to get a database of registered trademarks from the
country you are studying, then search for those marks in the corpus you are
examining. In the US, the database can be obtained from the U.S. Patent and
Trademark Office for about $300 per year.

-bruce

At 09:58 AM 11/1/99 +0000, Richard Evans wrote:
>Hello everyone,
>
>Does anyone know of any work relating to the automatic recognition of product
>names in documents such as technical manuals? I note that such names are not
>considered "markables" in the MUC-7 NER guidance, but it seems to me that
>recognition of product names (artifacts?) has something in common with
>named-entity recognition, in that both tasks require treatment of capitalised
>words. For example:
>
> "The TopTools package is included with the Omnibook, but you must
install it
>before you can use it. On Windows NT systems, you will need administrator
>privileges to install TopTools software."
>
>Directions to previous work and suggestions would be most appreciated.
>
>Many thanks in advance,
>
>Richard
>
>
>Richard Evans,
>Computational Linguistics Research Group,
>University of Wolverhampton,
>UK.
>