A Dynamic Semantic Framework to Support Multiple User Viewpoints during Information Retrieval

Semantic models offer powerful benefits to mediate between and to reason about heterogeneous data sources, when data from multiple data resources needs to be combined. A single domain conceptualisation and semantics can be projected as a single information viewpoint over multiple information sources, targeted to answer core queries by domain expert type users. This has the advantage that it avoids the necessity of users needing to be familiar with the stored data schema and associated query interfaces. However, some users may not be familiar with the full semantic model in order to pose and refine their queries under the strict constraints of a single conceptualisation. The single conceptualisation may also not adequately capture the operational nuances, side effects and errors in answering data queries. The development and application of a more flexible approach to derive more targeted views, adapted for different types of users and applications, is reported in this paper