Management of Dynamic and Imperfect User Context Information

In high-volume information systems, there is a trade-off between ease of querying and selectivity of the query (and thus relevance of the result set). This can be overcome by incorporating user context information into the interaction. But user context information is hard to acquire due to its dynamic nature and the approximative methods. This thesis work proposes a context model, which builds on annotations for qualifying and quantifying imperfection, and introduces the notion of sub contexts, which can be used to capture the dynamics of user context information. Its approach aligns with Semantic Web technologies.