dataxlogo.GIF (1005 bytes)             Automated Extraction of Business Rules
- - A Neurofuzzy Approach to Data Mining

(Click here for DataX™ Catalog)


Introduction to Business Rules
Business rules represent the intrinsic relationships between business entities. A business rule can be a relationship between gas price and local economy, or that between the spending capability and income level and professional occupation within a customer group. Credit companies and banks are faced with the big issue of fraud which costs them billions of dollars each year; they need to automatically detect fraud.  More important for them is the ability to automatically explain which and why people tend to fraud - an explanation process that is not available in most neural network-based data mining software.

Neuro-fuzzy Modeling
Neuro-fuzzy modeling is to use neural networks and fuzzy set theory to model practical systems. Zaptron has developed a proprietary neuro-fuzzy technology that can automatically extract business rules from the neuro-fuzzy model.  These rules can explain which and why people tend to fraud, or why a customer price increases.

A Case Study
One set of data is given below to demonstrate the rule extraction algorithm developed by Zaptron scientists.

(to be continued)

Questions and comments

Email to Last updated January 26, 1999
Copyright 1997-2000, ZAPTRON Systems, Inc.