We at the Publishing Group of the California Digital Library are working
on a project to automatically assign disciplinary subject terms to
content in our eScholarship Repostiory
(http://repositories.cdlib.org/escholarship/) as part of a larger
redesign effort. We are investigating classification tools right now,
so would be very interested to hear what others are doing and will share
the results of our work once we get something working. Also the folks
at UC Riverside Libraries are potentially investigating the use of
DataFountains with this same content, so hopefully there will be
information forthcoming about that effort.
Lisa
-----------------------------------------------
Lisa Schiff, Ph.D.
Technical Lead, Publishing
California Digital Library
300 Lakeside Drive #745
Kaiser Center
Oakland, CA 94612
510-987-0881 (t)
510-987-0243 (f)
www.cdlib.org
[log in to unmask]
-----Original Message-----
From: Repositories discussion list
[mailto:[log in to unmask]] On Behalf Of Julie Allinson
Sent: Wednesday, August 06, 2008 1:29 AM
To: [log in to unmask]
Subject: Re: Repositories using some form of automatically generated
metadata
Hi,
I know that Hull have been working with the data fountains software [1]
in their RepoMMan project [2] to extract metadata automatically by
analysing a document's contents and present this to users before asking
for direct metadata input. For York Digital Library where the focus is
on multi-media, we are hoping to explore a range of techniques,
including extracting information from deposited files, and offering a
range of pick lists and auto-completion options - we just haven't done
any of this yet. The AHDS MetaTools project is looking at creating some
prototype web services in this area too [3].
This is a very relevant topic so thanks for raising it. As someone very
interested in any work to reduce the metadata created by hand I think
there could be a very useful open debate on this on-list.
Cheers,
Julie
[1] http://datafountains.ucr.edu/
[2] http://www.hull.ac.uk/esig/repomman/
[3] http://www.ahds.ac.uk/about/projects/metatools/
Mahendra Mahey wrote:
> I am trying to find the extent to which repositories are using some
> form of automatically generated metadata.
>
> This could be in the form of automatically inserting the depositors
> details into the author field as a suggestion (if they are indeed the
> author - as sometimes they are not), a pick list appearing on a
> deposit form from an internal database, to the use of automatic
> classification systems that populate fields such as keywords, subject,
> title etc after an analysis of the item deposited.
>
> *Questions*
>
> If your repository is using auto metadata...
>
> What kind of auto metadata is being used and how? Has this been
> formally documented? Is this available? If not, could you provide me
> with a screnshot?
>
> If you are not using it, I am assuming that you would like to use some
> form of it, as long as it is reliable? If any of you are have
> objections or bad experiences to using auto generated metadata, please
> let me know why.
>
> Could you please *reply to me off list*?
>
> Thank you
>
--
Julie Allinson <[log in to unmask]>
Digital Library Manager
University Library & Archives, J.B. Morrell Library University of York,
Heslington, York, YO10 5DD, UK
tel: ++44 (0) 1904 434083 skype: j.allinson
web: http://www.york.ac.uk/services/library/elibrary/digitallibrary.htm
--
|