Semi-Automatic Discovery of Dependency Relationships Among Properties for Ranking-Based Semantics

Evaluation of argumentation involves judging the acceptability of presented arguments. For this, various semantics have been proposed as a ranking-based semantics that ranks arguments from the most to the least acceptable. At the same time, various properties have also been proposed to characterize and distinguish these semantics. However, studies focusing on the relationships among properties are insufficient, and knowledge about many properties has not been organized. Furthermore, there is no system in place to investigate them. Therefore, in this study, for investigating dependency relationships among properties, we construct a new system that implements a novel approach: by focusing on the absence of dependencies, it inversely reveals the relationships where dependencies do exist. We then present and discuss the obtained results.