Laud11Ch11

Mgt 4322 - Home Spring 2010 Email to Dr. Lyons PatLyons Home
[ Calendar10:10 | PortWebsite | StuWebs | Port1 | 2 | 3 | 4 | 5 | 6 | Showcase | Presentation ]
[ Ch 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 || AppDatabase | Ex1 | 2 | 3 | 4 | 5 | 6 | 7 ]
[ HW 1| 2 | 3 | 4 | | Career1| 2 | 3 ] [ SJU | TCB | CareerCtr | StuInfo | CareerLinks ] [ SJU Closing ] [H1N1SelfAssessment]

Ch 11 - Managing Knowledge

The Knowledge Mgt Landscape (p413)
1. Important Dimensions of Knowledge
  1. Def - Knowledge - the concepts, experience, and insight that provide a framework for creating, evaluating, and using information (page G7 of Glossary).
  2. Def - Tacit Knowledge - knowledge that resides in the minds of employees that has not been documented.
  3. Def - Explicit Knowledge - knowledge that has been documented.
2. Def - Knowledge Mgt - the set of processes (developed in an organization) to acquire, store, disseminate, and apply the firm's knowledge.
Enterprise-wide Knowledge Mgt Systems (p419)
1. Def - Structured Knowledge - knowledge that exists in formal documents, as well as, the formal rules used for decision-making.
2. Def - Semistructured Knowledge - all digital information that is not structured knowledge. This includes emails, voice mails.
3. Def - Enterprise Content Management System - system that classifies, organizes, and manages structured and semistructured knowledge. The system helps improve business processes and decisions throughout the enterprise. See Fig 11-4, p420.
4. Example - Enterprise Content Mgt System vendor - OpenText - www.opentext.com
  1. Central Vermont Public Service uses Open Text tool to help comply with governmental regulations.
  2. http://www.opentext.com/2/global/press-release-details.html?id=1858
5. Knowledge Network Systems
  1. Def - Knowledge Network System - system that provides an online directory of corporate experts in well-defined tacit knowledge domains.
  2. Example - Knowledge Network System vendor - AskMe - www.askmecorp.com, Fig 11-5, p421.
Knowledge Work Systems - skip (p424)
Intelligent Techniques (p427)
1. Capturing Knowledge: Expert Systems
  1. Def - Expert System - an information system that contains:
    1. Knowledge base - a file of rules and facts
    2. Inference engine - a computer program that examines existing rules and facts and infers new facts when possible.
      See http://en.wikipedia.org/wiki/Expert_system.
  2. Demonstration of Expert System Technology Assistant - in ESTA folder on S: drive.
    Part of book, Applying Expert System Technology to Business, by Patrick Lyons, published by Wadsworth Publishing Company, Belmont, CA, July, 1993.
  3. Examples of Successful Expert Systems
    1. CLUES - Contrywide's Loan Underwriting Expert System
      CLUES agrees with 95% of underwriters' decisions.
      Without CLUES, an underwriter handles 7 applications/day.
      With CLUES, an underwriter handles 16 applications/day.
    2. Authorizer's Assistant - developed for Travel Related Services division of American Express. It condensed 6" thick Authorization Manual to 600 rules. Resulted in less stressful job for authorizer (previously used 16 screens of data in less than 90 seconds) and improved consistency.
    3. Expert system vendor - Exsys - www.exsys.com - see www.exsys.com/case.html for case studies.
2. Organizational Intelligence: Case-based Reasoning
  1. Def - Case-based Reasoning System - a information system that contains:
    1. Database of past experiences of human specialists, stored as cases
    2. Pattern matching software that retrieves closely matching cases when presented with a new case.
  2. Case-based reasoning systems are helpful for diagnostic applications, such as medicine and customer support (help desks).
3. Neural Networks
  1. Def - Neural Network - a mathematical model that consists of:
    1. A number of simple processing elements, most commonly arranged in a 3 layer network, as shown in Fig 11-11, p434. Each processing element determines a single output value based on weighted sum of several input values.
    2. A training set of data. For the example in Fig 11-11, this data would contain input values for Age, Income, Purchase history, Frequency of purchases, and Average purchase size and the output value for Valid purchase and Fraudulent purchase for many cases.
    3. A training process. This process is used to adjust the weights so that the difference between the network outputs and training set outputs is reduced to a small value. See http://en.wikipedia.org/wiki/Neural_network.
  2. Neural network business applications:
    1. Detect credit card fraud
    2. Predict corporate bankruptcies
    3. Nonlinear regression. See http://www.patlyons.com/research/NeuralNets.htm.
4. Genetic Algorithms
  1. Def - Genetic Algorithm - a mathematical model that consists of:
    1. A genetic representation of all possible solutions, usually as an array of bits. See Fig 11-12, p437.
    2. A fitness function to evaluate all possible solutions.
    3. An evolutionary procedure, such as the following:
      Create an initial population of feasible individuals (solutions).
      Evaluate the fitness of each individual in the population.
      Select top-ranking individuals to reproduce.
      Breed a new generation through crossover (take substrings from each parent) and mutation (arbitrarily change a bit) and give birth to offspring.
      Evaluate the fitness of each new offspring.
      Replace bottom-ranking individuals with offspring.
      Repeat until an acceptable individual (solution) is found or computational limit is reached.
  2. Genetic algorithms find good feasible, not necessarily optimal, solutions. As a result, they are good for large problems that cannot be solved by optimization methods, such as scheduling problems with thousands of variables.
    
    Link to Chapter 11 outline with eyes.
    
    (This page was last edited on January 17, 2010 .)