Temporal Prediction: Dealing with Change and Interactions within a Causal Framework

This paper represents a first step in extending previous work on causation to deal with highly interacting patterns of change. A causal-temporal ontology is devised that is able to distinguish between different types of causation depending on the action-type of the entities involved in the causal rules and on persistence features of the phenomenon to be modeled. We introduce a discrete time structure and we associate appropriate temporal metric parameters to each causal rule in order to model delay and persistence. Moreover, several criteria have been developed for handling cases where different causal rules interact by predicting opposites effects on the same feature of the world. In the present paper we focus our attention on the task of temporal prediction and on the game between Ego and Nature.