Hybrid Approaches for Integrating Neural Networks and Case Based Reasoning: From Loosely Coupled to Tightly Coupled Models

Abstract

This chapter describes integration schemes for neuro-case based reasoning (neuro-CBR) systems. Three major schemes are distinguished: the neural networks can be used to implement a complete CBR system, to implement a special phase of the CBR cycle as retrieval or adaptation, and it can also be used separately with a CBR system in order to contribute to the accomplishment of a given task. We show example systems of the different schemes; we then present in detail a hybrid tightly coupled model for memory organization called ProBIS which was validated on different applications (diagnosis, control, classification).