Knowledge Bases by HC12070506945


									Knowledge Bases
Everything you ever wanted to know about
Knowledge Bases… You can say you learned here
from the VT-UVa Gurus!

                       Karen McDowell, Ph.D., GSEC
                       University of Virginia-Spring 2008
What is a Knowledge Base KB?
o Justan ole fashioned database for
  knowledge/information management
o Codifies resolutions to known problems
  and questions
o Means for computerized collection,
  organization, and retrieval of knowledge
o Machine-readable (~ semantic web/GKB)
  or human-readable (medical, technical)
o Home-grown or Vendor-supplied
What’s the Big Deal?
   Benefit from collective knowledge
   Transformation from linear and escalation based
    (streaming) to non-linear and collaboration-
    based (swarming) support
   Response to a business and cultural shift in the
    ways people use the Internet
   ROI unquestionably improves
   Context as important as text

Why the Interest in KBs?
   Users are more sophisticated and interact more
    with each other – collaborative
   Shift from mostly known to mostly new issues –
    tiered support not so effective
   99% of academic users now own a computer
    and use it all the time
   Internet is primary vehicle to solve problems –
    Users expect on-demand access
   Technical staff overwhelmed & student agents
    are transient
What Do We Want in KB?
   Impeccable Design
    ◦ Organize – Search – Inform
 Quality – Relevance – Accuracy
 Ease of Development - Low Maintenance
 Intuitive
    ◦ Works easily for internal and external users
Best Options with KB
   Context: Content filtered by context of environment,
    critical variables, & experience of intended audience
   Timeliness: Support knowledge available within hours
    of its capture (w/i ~ 3 hours)
   Choice of Access: FAQs by Product Area, Table of
    Contents, Lookup, Search Tool, Bookmarks, etc.
   No Dead Ends: Option for Assisted Support like Click
    to Call, Click to Chat and to Submit an Incident
   Feedback: Option to provide feedback and

What’s Good about KBs?
 Current information – consistent,
  efficient, and accurate
 Users like them – Better service
 Staff appear to be competent
 No need to reinvent the wheel
 Better ROI – reduces load on staff
What’s Bad about Them?
 Startup Intensive
 Maintenance
 Training – Data only as good as Input
 Fast pace of technological change
 More bloated? software to manage
What’s Available?
 RightAnswers
 PHP always a possibility
 University of Indiana code
 Wikis @ UVa
 Stock Answers @ MIT
 InSite @ Yale is dead six months later
 Perl freeware @ Princeton
Case Studies
 looks great – but…
  does it work? Our VT Guru will tell all!
 Awesome is made in
 Stanford & CUNY-Queens use
  RightAnswers –
 Duke uses Remedy and Chat
 Princeton freewares its OPM-Online
  Problem Management in Perl
Stanford uses RightAnswers

      But try and find this page from home page…very difficult -
Harvard: Eureka! HG KB

                  Kevin Davis SIGUCCS’02 /
Yale uses HG Internal KB

CalTech uses Google
MIT uses Stock Answers – HG

Princeton – HG KB Perl

UPENN HG? Many Dead Links

DUKE – Remedy + HG + Chat
Dartmouth HG Knowledge Library

University of Chicago / Web-based

UMICH / Single Page / No KB

CUNY – Queens – RightAnswers
eLibrarian - KB for a Library
 Overwhelmed by level of demand for reference
  services – used email
 Web-based eLibrarian 2.0 system KB fixed and
  improved system
 Increases the capacity of the service under
  constraint of a steady-state number of librarians
 Includes an auto-populated knowledge base and
  a recommender system - Anderson
Graduate School Management University of California at Los Angeles 2005
Brainstorming for KB Solutions
 We know from experience that 80% of what
  is in the KB will never be reused, 20% of what
  is in the KB will be reused, and some of that
  20% will be reused a lot.
 At the moment of creation we cannot know
  what the future value of the solution will be.
 Product evaluation against requirements list is
 Be aware of vendor assumptions! RightAnswers
  does nice work… IMHO
◦   A collection of data is not information.
◦   A collection of information is not knowledge.
◦   A collection of knowledge is not wisdom.
◦   A collection of wisdom is not truth.

            Fleming, Neil. Coping with a Revolution: Will the Internet Change
               Learning?, Lincoln University, Canterbury, New Zealand

To top