Constructing and Explaining Case Models: A Case-Based Argumentation Perspective
摘要
In this paper, we investigate constructing and explaining case models, which have been proposed as formal models for presumptive reasoning and evaluating arguments from cases. Recent research shows applications of case models and relationships between case models and other computational reasoning models. However, formal methods for constructing and explaining case models have not been investigated yet. Therefore, in this paper, we present methods for constructing and explaining case models based on the formalism of abstract argumentation for case-based reasoning (AA-CBR). The methods are illustrated in this paper with a legal example of paying penalties for a delivery company. With these two methods, we show an intended property that a dispute tree explaining the case model constructed from an AA-CBR case-base is homomorhpic to a dispute tree explaining the case-base itself. Additionally, we analyze that the methods are tractable in terms of number of cases and number of propositions used for representing each case.