A Semantic Approach for Learning Situation Detection

Abstract

A number of ubiquitous systems are considering how to deal with situation-awareness to act in a personalized manner according to user's needs. Indeed, most of the research efforts in situation-awareness are generally focused on the implementation of common characteristics, differing basically on the relevant context observed and how this contextual information is managed. Consequently, little attention has been paid to an abstract and general view for situation-awareness. The goal of this work is to provide a shift of the attention focus from the implementation to the modeling phase. Our objectives are (1) to define the conceptual modeling of situation-awareness for adaptive and ubiquitous learning systems, (2) to identify its relevant concepts, relationships, and structural, dynamic and representational elements, (3) to propose a method for situation detection. The domain concepts are explored by an application scenario that allows the understanding of the situations configured in u-learning systems and their dynamic factors. The conceptual model was implemented in the CONIC module, which was evaluated in an e-learning system specifying the application and constructing the improvements necessaries to deploy the system in a real application for tests.