A General Theory of Case Based Reasoning

Abstract

This chapter is the first exception to the Boolean structure of the book, although it is included in Part I of the book, as shown in the shaded area of Fig. 5.1. It is also the basis for Chapter 6 and 7, although this is not shown in Fig. 5.1. This chapter will attempt to provide a general theory of CBR, and move CBR towards a firm theoretical foundation based on similarity-based reasoning. To this end, this chapter will first extend the concept of similarity and examine similarity relations, fuzzy similarity relations, and similarity metrics. Then it provides a theoretical formalization for building case bases with three novel algorithms. It also proposes the R ⁵ model for CBR. Furthermore, it examines abductive CBR and deductive CBR and proposes a knowledge-based model for integrating abductive CBR and deductive CBR. Finally, it proposes rule-based models and fuzzy rule-based models for case retrieval.