A Semantic Approach to Data Integration from Clinical Polystore

Polystore refers to the data sources that span multiple data representations, structure and model. The data retrieval and integration systems experience challenges arise due to polystore. The Healthcare data retrieval applications combine the patient data from specialized clinical data stores. Resolving the heterogeneity of the data sources through transformation in single common data model is expensive and cannot be efficiently accomplished with Extract, Transform and Load (ETL)-based frameworks. The research work models the health data sources using Resource Description Framework (RDF) data standards to provides uniform access of underlying distributed data sources. The research work presents the ontology-based data integration solution with R2RML/RML mappings for relational and non-relational clinical polystores. The solution supports schema alignment which develops a common understanding of data. The work also presents the query translation of SPARQL to native database languages using R2RML/RML mappings to enable query answering across polystore data.