Training Intelligent Agents in the Semantic Web Era: The Golf Advisor Agent

Agent training techniques study methods to embed empirical, inductive knowledge representations into intelligent agents, in dynamic, recursive or semi-automated ways, expressed informs that can be used for agent reasoning. This paper investigates how data-driven rule-sets can be transcribed into ontologies, and how semantic web technologies as OWL can be used for representing inductive systems for agent decision-making. The method presented avoids the transliteration of data-driven knowledge into conventional if-then-else systems, rather demonstrates how inferencing through description logics and Semantic Web inference engines can be incorporated into the training process of agents that manipulate categorical and/or numerical data.