The Influence of Context on Intentional Service

Abstract

Abstract-Several service-oriented approaches promote the intention concept as a way to describe and document services based on user's requirements. However, these approaches have two main limitations: (1) they don't take into account the fact that a user evolves in a context that can influence his intentions, and (2) at the software service level, the corresponding intentional description of these software services is missing. Such a description should be a high level one, which is not directly connected to the software services. The objective of the paper is to propose a semantic service description that considers both intention corresponding to the service and context in which it is supposed to emerge. In addition, the variability embedded in the intentional description can be also affected by the user context. Such influence is also considered in our proposition.

A Context-Aware Intentional Service Prediction Mechanism in PIS

“…The intention (IU) is represented as an XML schema containing two mandatory elements, namely the verb and the target. Both verb and target are described by ontologies representing respectively significant actions made available by PIS and the objects considered by the actions [14]. The context (CxU) is also represented as an XML schema containing the context description.…”

“…Such descriptions follows an ontology-based context model on which context elements are described by an entity (corresponding to the entities whose context is observed) and a scope representing what is observed (location, memory, etc.) [14]. Finally, the service (Svi) represents the name of the service selected to satisfy this intention in this context.…”

“…Finally, the service (Svi) represents the name of the service selected to satisfy this intention in this context. Services are described using an extension of OWL-S we have proposed in [14], in which intention satisfied by the service and the context in which this intention is considered are described.…”

“…The clustering and prediction algorithms implementations are based on our OWL-S extended API [14], Jena [9] and the reasoner Pellet [18]. These impelentations use two classes, namely "ContextMatching" to determine the contextual matching score and the "IntentionMatching" class to determine the intentional matching score.…”

A user-centric vision of service-oriented Pervasive Information Systems

“…Similar to the discovery process, the semantic matching of intentions is based on ontologies describing these elements in order to calculate the score between them. Similarly, the matching between the user's context description and the context descriptions of the different states of the model is also based on a domain-specific ontology and on similarity measures between the values of context (see [26] for more details on the different ontologies).…”