SOFT-CBR: A Self-Optimizing Fuzzy Tool for Case-Based Reasoning
Kareem S. Aggour3
, Marc Pavese3
, Piero P. Bonissone3
and William E. Cheetham3 
| (3) |
GE Global Research One Research Circle, 12309 Niskayuna, NY |
Abstract
A generic Case-Based Reasoning tool has been designed, implemented, and successfully used in two distinct applications. SOFT-CBR
can be applied to a wide range of decision problems, independent of the underlying input case data and output decision space.
The tool supplements the traditional case base paradigm by incorporating Fuzzy Logic concepts in a flexible, extensible component-based
architecture. An Evolutionary Algorithm has also been incorporated into SOFT-CBR to facilitate the optimization and maintenance
of the system. SOFT-CBR relies on simple XML files for configuration, enabling its widespread use beyond the software development
community. SOFT-CBR has been used in an automated insurance underwriting system and a gas turbine diagnosis system.
References secured to subscribers.