A framework for ontological behavioral modeling in domestic dogs to predict aggression

There has been a wealth of study in the field of animal behavior; in the case of domestic animals, behavior is often studied in a shelter or laboratory environment [which is likely not predictive of domesticated dogs in the home]. Currently there are no standard methods with scientific merit to determine optimal, reliable, and consistent assessment techniques for domestic dogs and the behaviors that occur as a result of their situational contexts. Accurate behavior analysis is necessary both to reliably predict aggression and to study the effect of human interaction on innate behaviors. In this paper, we propose a contextual ontological framework by which to classify aggressive behaviors in domestic dogs by consulting with a panel of training experts. Further, we propose a statistical methodology for measuring aggression during conspecific [same species] socialization, and lastly we demonstrate this method on a small set of sample cases. Ultimately, we hope to apply this framework to a much larger set of cases to further develop the field and provide groundwork for making scientific the art of animal training.