SemiLD: mediator-based framework for keyword search over semi-structured and linked data

Abstract

Linked Data initiative has completely changed the procedure of sharing knowledge over the Web. It primarily aimed at improving the interoperability and semantics of the data published, by following a set of recommendations. Still, many data sources, which have a significant value, have not migrated to this new data space and continue to publish semi-structured data. Thus, new challenges arise in accessing and integrating the two data sources and models. This paper explores and identifies some of the major challenges, such as the continuous expansion and dynamism of a heterogeneous and an autonomous yet connected web of data, and addresses them by proposing SemiLD, a mediator-based framework to integrate on-the-fly heterogeneous semi-structured and Linked Data sources. The approach is implemented into a highly automated keyword search system that retrieves its input from various SPARQL endpoints and web APIs. The evaluation of the system illustrates the high precision, performance and recall of the contributed approach.