All,

Classification of unstructured electronically stored information (u-ESI)is
currently being addressed by a few vendors, including IBM, Kahn Consulting,
Autonomy and a few others.  Methods vary from the use of a taxonomy,  to the
use of 'knowledge models' comprised of examples of documents which are known
to belong to a particular retention category.  Performance is typically
validated using precision/recall metrics built into the software.  The
'knowledge model' approach leverages the fact that an organization knows
what types of content are intended to fall into a particular category, and
uses the full content of the examples for training.  This contrasts to a
taxonomy approach which relies on a list of key words and phrases, which
suffers from a very long list and problems with disambiguation between
categories.

I happen to work for Kahn, full disclosure.  Contact me offline if you would
like more education on this topic.  Here's a link to the Kahn website.

http://www.kahnconsultinginc.com/risk-management-shared-network-drive-clean-
up.php

Brent Stanley 

-----Original Message-----
From: Records Management Program [mailto:[log in to unmask]] On Behalf
Of Olsen, Christopher J.
Sent: Tuesday, May 11, 2010 5:05 PM
To: [log in to unmask]
Subject: Examples of Auto-classification of records in private and/or public
organizations

To all:

As background, in the mid-90's the community was all aflutter with the
promise of auto-classification technologies-but, there were few examples of
any large-scale use.  The vendor community pretty much dried up and those in
it are now in search technologies.  With the few exceptions of vendors
offering a rudimentary "auto-classification" capability with their records
management applications, it seems to be little in the space.  No mention of
significant efforts by ARMA, AIIM, Cohasset, etc.

Admittedly, I have not done exhaustive research on the web but, I have not
found any significant reporting by public or private organizations about
their use of auto-classification tools, either as a feature of a records
management application, or as a s/w product installed on the content
management system.

If you have seen or heard of any successful auto-classification efforts,
especially using the organization's own taxonomy as the rules engine, I
would appreciate getting the URL or information.  I suspect there may be
some pilot efforts going on and that would be good to know.  But, learning
of examples of an enterprise use in an organization with thousands of people
and a somewhat complex file plan would be so welcome.

Thanks,

Chris


"You are making PROGRESS if, each mistake you make is a NEW one?"
Christopher J. Olsen, CRM
Systems Engineer, Principal
Lockheed Martin
SENTINEL Program
Office:  703-483-4134
Fax:    703-889-1871
Cell:    703-340-5817
email:  [log in to unmask]


List archives at http://lists.ufl.edu/archives/recmgmt-l.html
Contact [log in to unmask] for assistance To unsubscribe from
this list, click the below link. If not already present, place UNSUBSCRIBE
RECMGMT-L or UNSUB RECMGMT-L in the body of the message.
mailto:[log in to unmask]

List archives at http://lists.ufl.edu/archives/recmgmt-l.html
Contact [log in to unmask] for assistance
To unsubscribe from this list, click the below link. If not already present, place UNSUBSCRIBE RECMGMT-L or UNSUB RECMGMT-L in the body of the message.
mailto:[log in to unmask]