Pay-As-You-Go Gathering of Domain Knowledge for Information Quality and Integration
There are many data manipulation tasks that require a knowledge of domain semantics to be carried out effectively. For example, finding and resolving data quality problems typically requires an understanding of the semantics of the data and the context of both its creation and use.
Similarly, the integration of data from different sources is often dependent on an understanding of the underlying data semantics and current conventions for representation of data. Until recently, we have depended on the availability of teams of willing domain experts and technical experts working together to extract the required semantics and to encode them in some usable form. However, the high costs of this approach (coupled with the short shelf-life of the results in many cases) have lead us to explore alternative approaches. In this project, we will continue current work in the school on techniques and tools for "pay-as-you-go" gathering of domain semantics, focussing in particular on mechanisms by which useful domain semantics can be gathered from domain experts without technical knowledge in a cost-effective manner.