Semantic Web Technologies in Health Care Analytics
Over the past years, the KRR group around Ian Horrocks in the Computer Science department has developed RDFox, an optimised RDF-triple store and parallel Datalog reasoner. The reasoner's high performance allows RDFox to handle large amounts of data with complex Datalog programs. The usefulness of this technology and its applicability in the real world were demonstrated in the project whose results are to be presented in this talk. The project was partly funded by the U.S. health care provider Kaiser Permanente and the DBOnto platform grant. The aim of the project was to compute health care quality measures issued by the National Committee of Quality Assurance (NCQA) in the U.S. on real data from one of the regional branches of Kaiser Permanente. The project has been successfully completed and the talk will discuss
- the principles underlying RDFox's technique of data derivation
- the data model that was used in order to represent the provided data
- examples of Datalog rules encoding passages of the given NCQA specification
- computation times
- how the project shaped the future road map of RDFox development
The talk is self contained and no presupposed knowledge is needed.