Hüllermeier E. Case-Based Approximate Reasoning

Dordrecht: Springer. − 2007. – 382 p. In the last three decades, important trends of research in artificial intelligence have been devoted to the design and the study of inference systems that exhibit humanlike reasoning capabilities, with special emphasis on tolerance toward incomplete information and uncertainty. This program, which can be also related to cognitive psychology concerns, has led to the development of computational models, in particular for default reasoning that accommodates exceptions and inconsistency, for approximate reasoning with interpolative abilities, as well as for case-based reasoning. Case-based reasoning (CBR) relies on the concept of similarity, and more particularly on the idea that situations recognized as similar in important aspects may be also similar in other respects. CBR thus appears as a yet simple instance of analogical reasoning, but as powerful as a general problem solving method. This explains the success encountered by case-based reasoning, and beyond that, the interest for similarity-based reasoning that has been increasing in the last ten years. The very idea of CBR is thus to solve new problems on the basis of experience that is represented by already solved problems of the same type, referred to as cases. Thus, a new problem is solved by adapting the solution of a similar case, hoping that the adaptation can be done with much less effort than solving the problem from scratch.Similarity and Case-Based Inference
Constraint-Based Modeling of Case-Based Inference
Probabilistic Modeling of Case-Based Inference
Fuzzy Set-Based Modeling of Case-Based Inference I
Fuzzy Set-Based Modeling of Case-Based Inference II
Case-Based Decision Making
Conclusions and Outlook
Possibilistic Dominance in Qualitative Decisions
Implication-Based Fuzzy Rules as Randomized Gradual Rules
Similarity-Based Reasoning as Logical Inference
Simulation Results of Section 3.4.4
Experimental Results of Section 5.5.4
Simulation Results of Section 7.4
Computation of an Extended Splitting Measures
Experimental Results of Section 7.7.2