"This paper presents a semantic network of commonsense knowledge that is useful in
capturing human experience in ambient intelligence environments. Human behavior is
composed of a series of behavior patterns. And a behavior pattern contains behavioral
features, such as the objects touched by users, time, and location information, within a
period of time. Consequently, in this paper, the algorithm of the spreading activation
model is adopted. Each behavior node is inferred based on the possibility of what
behavior may happen next and on their mutual relationships. In the spreading
activation model, various weights are put into certain nodes and checked to see if
some behavior nodes are enabled under such a circumstance."
