A Cluster-based Approach to Web Adaptation in Context-Aware Applications

In mobile Web applications, there is the need to take into account context characteristics (such as the device capabilities, the network QoS, the user preferences, and the location) to meet constraints of the client and guarantee a satisfying interaction with the user. A major issue in this framework is that, in real world scenarios, the number of adaptation requirements can change and increase very rapidly. Therefore, a relevant problem is the definition of effective methods for choosing efficiently the most suitable adaptation for a given context. To this aim, we propose in this paper a new cluster-based approach that automatically classifies the the contexts on the basis of their characteristics: at a logical level, each class corresponds to contexts that require similar adaptations. We show that this classification strongly alleviates the adaptation process. The approach relies on a metric distance that is used to compare contexts and on a threshold that is used to group them. Each context in a cluster is associated with the adaptation that best matches with the context requirements. We also illustrate an implementation of our approach and a number of experimental results that support its effectiveness. Semantic Web design principles and enabling technologies are important ingredients of the overall framework.