Volume 12, Issue 5
|
|
Volume 12, Issue 5
|
|
|
|
| Sept/Oct 1998 | |
| Theme: Data Mining, Genetic Algorithms, and Modeling |
|
To Volume 12, Issue 4
|
To Volume 12, Issue 6
|
||||||||
 |
 |
||||||||
|
Features
|
| Knowledge Discovery and Data Mining: The Expectation of Magic - Dorian Pyle describes a methodology to derive the most benefit from this technology which involves participation by people throughout an organization. |
| Finding Robust & Usable Models with Data Mining: Examples from Finance -- Vasant Dhar and Roger Stein show how to use data to build effective models using two common applications, credit analysis and securities trading. |
| Data Mining: Issuing Predictions and Revealing Unexpected Phenomena -- Abraham Meidan provides a solid introduction to this technology and its capabilities. |
| Knowledge Management and Microsoft's Active Platform: Part 2 - Henry Seiler outlines the construction of an Internet-integrated client/server application in this conclusion to his series. |
| Book Review: Solving Data Mining Problems through Pattern Recognition, Predective Data Mining, and DMSK- Will Dwinnell sifts through the available data mining books & reports on two exceptional ones, plus a corresponding software package. |
| AI@Work - Salford Systems, Palisade Corporation, and IBM share customer success stories from data mining for better results with Cabela's catalog mailing, to genetic algorithms helping John Deere improve production, to data mining for better customer relations and profit with the Bank of Montreal. |
| Regulars | ||
| Editorial | ||
| Secret Agent Man - Build Your Own Chatterbot with NeuroStudio | by Don Barker | |
| Intelligence Files - Hot Fun in the Summertime | by David Blanchard | |
| AI and the Net - Semantic Nets on the Net and the coming of X | by Mary Kroening | |
| Product Updates ---------------------------> | 14 late breaking product announcements from around the world in the fields of: | |
| Announcement | Data Mining | |
| Expert System Development Tools | Fuzzy Logic | |
| Internet | Languages | |
| PC AI Buyer's Guide -----------------------> | Consulting | Data Mining |
| Genetic Algorithms | Modeling | |
| Training | ||
| Product Service Guide - Provides access to information on an entire category of products | ||
| PC AI Blackboard - AI advertisers bulletin board | ||
|
Advertiser List for 12.5
|
||
| AAAI | Intelligent Systems Report | Prolog Development Ctr |
| AI Developers Inc. | Intellix A/S | Rule Machines Corp |
| Amzi! Inc. | ISoft | Salford Systems |
| Applied Logic Systems | Knowledge Tech. E-Commerce | Soft Warehouse Inc. |
| ATTAR Software USA | Knowledge Tech Know. Mgmt | Sonalysts Inc. |
| BotSpot | Logic Programming Assoc Ltd | Stellar Technology |
| California Scientific Software | Morgan Kaufmann | System Dynamics Intl |
| COMDEX/Enterprise | MIT GmbH | The Haley Enterprise |
| DCI | Nestor Inc. | The MathWorks |
| Franz Inc. | Neuron Data | Ultrexx Corporation |
| Frontier GlobalCenter | NeuroDimension | Ward Sys. Group Inc. |
| Gordian Institute | Palisade Corporation | WizSoft Inc |
| Harlequin . | Production Sys Tech Inc. | Xi Computer Corp |
| IBM | Prentice Hall | |
| IJCNN '99 | Production Sys Tech |
| All businesses are concerned about the bottom line. Yet less than 10% use AI technology to discover crucial information they already have - on their computers. In the next five years, these companies will face increased challenges and opportunities provided by the competition of a global economy. These advance technologies will be a competitive resource that differentiates the more profitable from the less profitable companies. Commercial grade AI-based information technology has been available for years and has a great track record. | |
| While searching the web recently, I found hundreds of success stories. For example, major retailers such as Sears are mining data from their millions of credit card holders - grouping them into buying patterns for targeted promotions and special marketing programs. Another giant retailer wanted to identify its most frequently linked purchases. Using AI techniques to search its database, it found the unlikely association of beer an ddiapers. When placed next to each other in the stores, sales of both increased. Safeway supermarkets carried over two dozen brands of orange juidce until they identified the eight best-sellers through AI technology. They now benefit from increased sales as well as reduced floor space, creating room for other products. MCI uses similar techniques to detect when a customer is about to leave. Once identified, MCI initiates marketing efforts to retain them - you may have experienced this first hand. McDonald's analyzes its menus, sales, profitability, and combinations of products at various price points to develop reliable predictive models. The uses of AI technology are as diverse as the available data: automobile owner demographic marketing data; clinical hospital patient forecasting; business fraud detection; stock market prediction; production quality control; banking and insurance industry risk analysis, and even agriculture. | |
| In this issue we focus on three AI technologies: data mining, modeling and simulation, and genetic algorithms. Data mining reveals the tactical and strategic information hidden in large databases. Many of the more publicized success stories involve data mining. Modeling and simulation tools enable organizations to diagram and analyze their current and future business and technical processes. Also improved are consensus building; documentation; gap and overlap analysis; optimization; plus they provide a connection to application modeling, design, construction, and total systems monitoring. Genetic algorithms, in conjunction with technologies such as neural networks, expert systems, fuzzy logic, simulated annealing, classical search, and optimization are transitioning from the research arena into commercial applications. This field has a lively Usenet group (see comp.ai.genetic). | |
| This issue provides a technical look at data mining from methodology and data partitioning to predictions. Abraham Meidan provides an overview of the technology in "Data Mining: Issuing Predictions and Revealing Unexpected Phenomena." Dorian Pyle defines a structured methodology that can lead to successful deployment in the article "Knowledge Discovery an dData Mining: The Expectation of Magic." In "Finding Robust & Usable Models with Data Mining: Examples from Finance," noted authors Vasant Dhar and Roger Stein provide a generic framework for partitioning data to generate effective models. To round out the theme, Will Dwinnell, a regular contributor to PC AI Magazine (and new father), reviews two excellent data mining books and a corresponding software package. | |
| In our first AI@Work application, a hunting, fishing, and outdoor gear outfitter uses Salford Systems' data mining techniques to improve its catalog mailing performance models. In the next story, John Deere uses Evolver, a genetic algorithm-based package from Palisade Corporation, to improve factory scheduling and processing. To conclude, we see how IBM and Bank of Montreal implemented a database marketing and customer relationship management system. | |
| Also in this issue, Henry Seiler wraps up his series on knowledge management in "Knowledge Management and Microsoft's Active Platform: Part 2," Mary Kroening, in the "AI and the Net" column covers the use of semantic nets to define relationships between objects and how this might be used with XML (Extensible Markup Language). Our agent Don Barker reveals the secrets behind constructing a bot in his "Secret Agent Man" column. Dave Blanchard, industry watcher, updates us on the latest AI related acquisitions and partnerships in his "Intelligence Files" column. | |
| As the world gets smaller, through advances in the global economy and e-commerce, it's going to get more and more competitive out there. Businesses large and small will benefit from knowing what draws and keeps their customers, and AI technology is ready and waiting for the opportunity to "show them the money." | |
| Terry Hengl |
|
||||
|
|
|
|
|