I. Introduction
THE present work generalizes and extends a method proposed in [2] to find event patterns. We consider events having a twofold structure: on one hand, they happen in a temporal and, when deemed important, spatial dimension, so that they are immersed in a continuous euclidean space; on the other, they involve entities from some domain. The continuous space where events are immersed is this way a means to put them in relationship and develop reasoning. We note that, in specific cases, continuous properties other than time and space may be relevant, for instance, amounts of money transfer; making the proposed method generalizable. The actual graph representation may vary depending on the aims of the research and the employed algorithms: events can be considered edges linking entities or nodes sharing edges when they have entities in common; sometimes a mixed strategy can be used.