Browse Prior Art Database

Training a Context Database Using Expert Knowledge

IP.com Disclosure Number: IPCOM000105381D
Original Publication Date: 1993-Jul-01
Included in the Prior Art Database: 2005-Mar-19
Document File: 2 page(s) / 38K

Publishing Venue

IBM

Related People

Kevern, B: AUTHOR

Abstract

A facility that allows a context database to be trained with expert knowledge such that the context database can tell a novice user what command an expert would enter next given the past sequences of entered commands is disclosed. The power of this facility comes from the ability to easily capture the knowledge of an expert in some facet of using the operating system and to make this knowledge available to a novice user such that they can quickly learn what the expert would do in a particular situation.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 87% of the total text.

Training a Context Database Using Expert Knowledge

      A facility that allows a context database to be trained with
expert knowledge such that the context database can tell a novice
user what command an expert would enter next given the past sequences
of entered commands is disclosed.  The power of this facility comes
from the ability to easily capture the knowledge of an expert in some
facet of using the operating system and to make this knowledge
available to a novice user such that they can quickly learn what the
expert would do in a particular situation.

      A user class is defined as a class of users executing similar
tasks on the system.  By classifying a user as a member of a
particular class, the users within that user class can make use of
the context database that contains the contexts typical of users
within that class.

      Not only can the typical contexts be added to this database,
but the contexts from an interaction with an expert within the user
class can be used to train the context database.  By doing this, all
users of the context database for a particular user class can see
what an expert would do next in a given situation.

      Gathering expert knowledge to train the context database is as
simple as letting the expert perform their job on the system and then
capturing the job log of that expert.  Once captured, the context
database mechanisms provided allow the job logs to be fed in as
training data to the context database....