Enabling proactiveness through context prediction

Abstract. In order to enable proactiveness in applications and services, inferences about likely future contexts are needed. This task, which is often referred to as context prediction, has not been properly addressed in the literature. We define the different components of context awareness from data acquisition to prediction. Some mathematical principles of making predictions, alongside with technical challenges that must be solved in systems producing predictions about context are discussed. Several possible uses for predicted context are presented as well as some architecture issues related to the implementation of context prediction. 1