Dynamic rule mining for argumentation based systems

Argumentation has proved to be a very influential reasoning mechanism particularly in the context of multi agent systems. In this paper we introduce PADUA (Protocol for Argumentation Dialogue Using Association Rules), a novel argumentation formalism that dynamically mines Association Rules (ARs) from the case background as a means to: (i) generate the arguments exchanged among dialogue participants, and (ii) represent each participant's background domain knowledge, thus avoiding the traditional knowledge base representations. Dialogue participants mine ARs from their own case data and then use these rules as arguments and counter arguments. This paper fully describes the PADUA formalism and proposes a suite of dynamic ARM algorithms to provide support for the argumentation process.