Combining nominal and continuous properties in an incremental learning system for design

Research in machine learning has produced many learning algorithms mainly for classification tasks. This paper reports on an extension made to the learning program COBWEB to allow it to handle examples described by a more complex description language. The paper describes Bridger, a system that implements this, as well as other extensions. Bridger, combines leaning and performance in design tasks. The extension implemented has been tested succesfully in four design domains. This extension and others are necessary for allowing Bridger to master design rather than simple classification tasks. This work has supported in part by the Engineering Design Research Center, a National Science Foundation Engineering Research Center, and the Sun Company Grant for Engineering Design Research. UNIVERSITY LIBRARIES CARNEG32 FLL . ' : " T ^