Where Intelligent Technology Meets the Real World
Site Contents
Contact PC AI
Volume 15, Issue 5
Also check out the PC AI Article Summary List for the past 16 years
Sept/Oct 2001
Theme: Data Mining, Modeling, Simulation and Analysis

To Volume 15, Issue 4
Order Back Issues
To Volume 15, Issue 6

Building Intelligent Models from Data Mining and Expert Knowledge: A Look at Fundamental Principles Earl Cox explores key issues associated with knoweldge-based model design and execution including data variables, data and fuzzy spaces, experimental controls, coping with noise, ambiguity, missing data, isolation of dependent and independent variables, and statistical and regression analyses.
Data Mining Using Sample Data: A Sample based Modeling Strategy James Morgan, Robert Dougherty, Allan Hilchie, and Bern Carey illustrate how the use of data sampling improves model accuracy while lowering development and maintenance cost.
AI@Work Using data mining to prevent customer attrition.
Fueling the Search Engine with Natural Language: Developing with Smalltalk Robert Ingria and Dave Ginter detail the use of Smalltalk to develop a search engine that does not rely on Boolean logic.
Profitability and Mining Web Data: Avoiding the Path to Red Ink Scott Clendaniel discusses how a well-intentioned focus on customer response rates and similar dependent variables may cause devastating results and some techniques to avoid these pitfalls.
Text Processing in an Integrated Development Environment (IDE): Integrating Natural Language Processing (NLP) Techniques Paul Deane, David de Hilster, Amnon Meyers explore a new breed of IDEs that integrate NLP techniques with basic text analysis tasks - enhancing productivity when developing difficult text analysis applications.
Tools that Deal with Big Data: Modeling and Analysis Will Dwinnell demonstrates how a corporation's ability to analyze and model large data sets is crucial to a corporation's success.
Expertise through Rules Based Representation: Why Can't a Computer Program Write Computer Programs Glenn Hofford explores a rule-based approach to programming and software development that uses artificial intelligence to write programs.

AI and the Net - Virtual People: Final Fantasy by Mary Kroening
Product Updates ------------------------------> 22 late breaking product announcements from around the world in the fields of:
  Custom Service Update Data Mining
  Help Desk Intelligent Agents
  Intelligent Tools Internet and Web
  Modeling and Simulation Natural Language Processing
  Neural Networks Announcements
Product Service Guide - Provides access to information on an entire category of products    
PC AI Blackboard - AI advertisers bulletin board    

Advertiser List for 15.5
AAAI  Franz Inc. RML 
Amzi! Inc Frontline Rule Automation
And Corporation LPA Rule Machines Corp.
ATTAR Software USA Megaputer Intelligence Salford Systems
Business Rules Forum  NeuralWare StatSoft
CDI NeuroDimension TAI Inc
DCI  PC AI Banner Ad The Haley Enterprise
dtSearch PDC The Modeling Agency
Expert Reasoning Systems Production Systems Tech WizSoft Inc. 
Exsys QMC 


Knowledge Discovery

Data Mining as well as simulation and modeling are part of a much larger process - Knowledge Discovery in Data (KDD). Data mining, in a simple sense, is finding information hidden within a collection of data. While the relationships and patterns used to create the data are known, it is the relationships and patterns that are not known that potentially offer the real value. Modeling and simulation, on the otehr hand, often involve the exploration and discovery of concepts, ideas, and even knowledge, possibly using data as either a starting point or a form of model verification.
  Driving the need for new applications and new technology, the KDD process often groups data into three focus areas: -Business and Financial Data -Web Related DAta -Scientific Engineering Data.
  Examining these areas as well as a number of others, this issue delivers insight into some of the latest enhancements to the KDD process.
  Starting with the Business and Financial Data, we asked the Center for Data Insight, at Northern Arizona University to share the results of their studies involveing the data mining of large corporate data sets. In this issue, Earl Cox continues his series on knowledge-based model design and combines many of the techniques discussed in his previous articles, giving the reader a more complete picture of the model design process. He ties together dependent and independent data variables, data and fuzzy spaces, coping with noise, ambiguity, and missing data, variables, and statistical and regression analyses. He also illustrates how subject matter experts and the design and development of a conventional rule-based expert system are an integral part of the modeling methodology. A side note: Earl is writing a historical perspective on the development and deployment of AI technologies and applications used in business and industry, over the last 20 years. He is looking for annotates, case studies, successfully deployed applications, and interesting milestones. He can be contacted at earl.cox@panacya.com.
  Moving on to Web Related Data, Scott Clendaniel looks at the relationshps between profitability and data mining web data. He has discovered that a focus on customer response and similar dependent varibales often causes devastating reductions in profitability instead of the expected increases in earnings. Rounding out the three focus areas, Will Dwinnell finds tools to assist in dealing with large quantities of data. Although originally more of a problem for scientific research, processing large data sets is becoming a common problem for corporations.
  There are a number of sites providing solid information on data mining. A few that I discovered while preparing for this issue include Visualizing Data Mining Models (www3.shore.net/~kth/text/dmviz/modelviz.shtml), which includes an overview, academic papers, tutorial and index; and ADC's Data Mining Resources for Space Science, a useful NASA site (adc.gsfc.nasa.gov/adc/adc_datamining.html).
  Also in this issue, Glenn Hofford explores a subject of great interest to me - the concept of computers writing computer programs. This is the concept that originally captured my interest in AI. Robert Ingria and Dave Ginter discuss the use of Smalltalk to develop a Natural Language front end for search engines. This creates a search engine that does not require an understanding of Boolean Logic. Mary Kroening introduces us to Ramona and the concept of virtual people.
  Our Webmaster, Ilana Marks, has given a much needed face-lift and upgrate to the PC AI web site (www.pcai.com). Besides giving it a more modern look and greatly simplifying informal browsing within the site, she has added numerous helpful features. For example, a dynamic search-engine table under each technical category examines key search engines for the general category - providing links to the latest related information.
  We have started a moderated discussion group for sharing information on the topic of artificial intelligence. Subscribers to this list can share information with other subscribers and there will be postings of AI related news items from time to time to help initiate discussions. Although lightly moderated to keep out spam and keep within the general categories of AI, PC AI will not control the threads. You can visit http://groups.yahoo.com/group/pcai for information on how to subscribe to this free service or simply send an e-mail message to pcai-subscribe@yahoogroups.com.

Terry Hengl

Volume 15----------------------> Issue 1 (Jan/Feb 2001)   Volume 15 Index (2001)
  Issue 2 (Mar/Apr 2001)   Volume 14 Index (2000)
Issue 3 (May/Jun 2001)   Volume 13 Index (1999)
Issue 4 (Jul/Aug 2001)   Volume 12 Index (1998)
Issue 5 (Sep/Oct 2001)   Volume 11 Index (1997)
Issue 6 (Nov/Dec 2001)   Volume 10 Index (1996)
      Volume 9 Index (1995)
      Volume 8 Index (1994)

PC AI Magazine

PCAI.TH @ gmail.com
Copyright (C) 2001-2002, All rights reserved
Designed by Terry H. and the web team at Knowledge Technology: Comments? webmaster @ pcai.com