Brought to you by EarthWeb
IT Library Logo

Click Here!
Click Here!

Search the site:
 
EXPERT SEARCH -----
Programming Languages
Databases
Security
Web Services
Network Services
Middleware
Components
Operating Systems
User Interfaces
Groupware & Collaboration
Content Management
Productivity Applications
Hardware
Fun & Games

EarthWeb Direct EarthWeb Direct Fatbrain Auctions Support Source Answers

EarthWeb sites
Crossnodes
Datamation
Developer.com
DICE
EarthWeb.com
EarthWeb Direct
ERP Hub
Gamelan
GoCertify.com
HTMLGoodies
Intranet Journal
IT Knowledge
IT Library
JavaGoodies
JARS
JavaScripts.com
open source IT
RoadCoders
Y2K Info

Previous Table of Contents Next


5.1. CASE-BASED REASONING (CBR)

CBR relies on expertise in the form of worked cases. CBR works by measuring how similar a new situation is to an existing case in the case base, and retrieving the most similar (most relevant) case. A case consists of the attributes and values of a problem situation, as well as its solution. CBR does not require much of a domain theory -- it requires nothing more than a representative sample of stereotypical worked cases. Knowledge discovery and database mining techniques can often generalize, induce, and transform cases, or portions of cases, into useful rules.

Case-based reasoning should be used when:

  • The solution alternatives can be explictly enumerated
  • Numerous examples of worked cases exist that cover domain knowledge
  • No domain model or theory exists
  • Experts represent domain in terms of cases
  • Human experts or expertise are not available
  • Situational information is conflicting, uncertain, or missing
  • Knowledge is volatile and dynamic
  • Domain knowledge and expertise already captured by past cases
  • Workforce experience and performance are low
  • Need a fast way to acquire domain knowledge
  • Want to illustrate an outcome or explanation with an example
  • Want to contrast and compare potential solutions
  • Want to assess pros and cons of a situation
  • Want to test a theory or solution with cases

5.2. RULE-BASED SYSTEMS (RBS)

RBS use many small slivers of knowledge organized into conditional If-Then rules. Inference engines for RBS are either goal-driven, backward-chaining, or data-driven, forward-chaining, depending on the type of application or generic task. Rules often represent heuristics --shortcuts or rules-of-thumb for solving problems. Regarding the amount of structure, RBS fall in between CBR and MBR -- rules of thumb are abstracted and generalized from experience into small chunks of knowledge. Both CBR and RBS can be developed incrementally and can provide some value in an unfinished state. RBS can also be transformed into objects or frames using knowledge discovery and database mining techniques.

Rule-based systems should be used when:

  • Human experts think in terms of rules and heuristics
  • Task involves decision-making, problem-solving, heuristics, or judgment
  • Domain is complex and substantial expertise exists
  • Knowledge is stable and is well- or semi-structured
  • Expertise is primarily symbolic, not numeric
  • Human experts are willing and available for knowledge elicitation
  • Work performance and product quality are poor
  • Employee turnover is high and training is expensive
  • Impending loss of domain experts

5.3. MODEL-BASED REASONING (MBR)

MBR provides a representational and conceptual framework for knowledge that defines both knowledge structures and inferencing methods. MBR defines and structures relevant domain objects/concepts, their attributes, and their behaviors in order to organize work in complex domains and perform simulations. MBR also defines the relationships between the objects in terms of class hierarchies, composition, and causation. The most basic knowledge structure, governing all types of knowledge, is the <Object Attribute Value> triple.

MBR encompasses, represents, and organizes all types of knowledge, including CBR and RBS, as well as databases, text, images, and other media. Types of MBR are object-based, frame-based, and domain-specific; MBR models can also be categorized as quantitative or qualitative. MBR requires a well-structured, well-understood domain theory. In simulation, MBR components are often so tightly linked together that MBR has limited value without a completed model. MBR is also very useful for organizing and structuring complex domains and work processes.

Model-based reasoning should be used when:

  • A consensus framework of concepts and domain theory exists
  • Business processes, methods, and events need to be represented and modeled
  • Want to represent and organize large-scale, complex systems
  • Want to simulate performance and side-effects from future work system design
  • Want to control, monitor, and measure information workflows
  • Want to represent, organize, and integrate elements of knowledge repositories and related performance support systems
  • System navigation and presentation are important
  • Environment and data are relatively dynamic
  • IT infrastructure uses a client/server architecture
  • Results from knowledge elicitation and acquisition need to be organized

6. APPROACHES TO IDENTIFYING EXPERT SYSTEM APPLICATIONS

Following is a list of approaches in which potential ES applications have been found:

  • Knowledge-Intensive Areas:
    • Industries
    • Organizational functions
    • Work activities
  • Business system components
  • ES generic task
  • ES typology: when to use which type of ES (covered in Section 5)

6.1. KNOWLEDGE-INTENSIVE ORGANIZATIONAL FUNCTIONS APPROACH

One key to finding promising ES applications is to locate industries, organizational functions, business activities, and work system components where work is information- and knowledge-intensive.

The following organizational functions are information- and knowledge-intensive:

  • Research and Development
  • Engineering
  • Marketing
  • Sales
  • Customer Service
  • Management
  • Contracting
  • Legal
  • Human Resources/Administration
  • Training
  • Information Systems

6.2. KNOWLEDGE-INTENSIVE WORK ACTIVITIES APPROACH

On a finer-grained level within an organization, activities within functions can also be classified as information- and knowledge-intensive. These following business activities that are knowledge-intensive can occur within any of the organizational functions:

  • Communicating
  • Facilitating, coordinating, negotiating
  • Managing
    • Planning
    • Budgeting
    • Organizing
    • Controlling
    • Evaluating
  • Creating/retrieving/updating/distributing information
  • Decision-making
  • Problem-solving
  • Researching
  • Designing
  • Teaching, coaching, tutoring, training
  • Learning
  • Diagnosing, repairing
  • Computing
  • Perceiving, sensing


Previous Table of Contents Next

footer nav
Use of this site is subject certain Terms & Conditions.
Copyright (c) 1996-1999 EarthWeb, Inc.. All rights reserved. Reproduction in whole or in part in any form or medium without express written permission of EarthWeb is prohibited. Please read our privacy policy for details.