Service-centric Inference and Utilization of Confidence on Context

The inadequate quality of context forces the context consumers in pervasive environments to reason about the quality and relevance of context to be confident of its worth to perform their functionality. The additional task of analyzing large volumes of context drastically affects the performance of the context consumers to adjust to dynamically changing situations. A single value that presents the quality and relevance of context information tailored to the needs of a particular context consumer may release them from spending resources on context quality analysis and let them concentrate on their main task. In this paper we present a novel technique to combine different Quality of Context (QoC) metrics to infer the value of confidence on context. Our technique also considers the requirements of a particular context consumer regarding QoC metrics while confidence inference. Confidence on context is further provided to the context consumers to select high quality context and use the confidence in their functionality. We have successfully evaluated our approach using two context consumer services and user context collected from a smart home pervasive environment.