An experimental analysis on the similarity of argumentation semantics

In this paper we ask whether approximation for abstract argumentation is useful in practice, and in particular whether reasoning with grounded semantics - which has polynomial runtime - is already an approximation approach sufficient for several practical purposes. While it is clear from theoretical results that reasoning with grounded semantics is different from, for example, skeptical reasoning with preferred semantics, we investigate how significant this difference is in actual argumentation frameworks. As it turns out, in many graphs models, reasoning with grounded semantics actually approximates reasoning with other semantics almost perfectly. An algorithm for grounded reasoning is thus a conceptually simple approximation algorithm that not only does not need a learning phase - like recent approaches - but also approximates well - in practice - several decision problems associated to other semantics.