Feedback-based data set recommendation for building linked data applications

The huge and growing volume of linked data is increasing the interest in developing applications on top of such data. One of the distinguishing features of linked data applications is that the data could come from any RDF data set available on the Web. Different from conventional applications, where the data sources are under control of the application's owner or developer, linked data applications follow the Semantic Web vision of a world full of reusable data. Considering a potentially large number of data sets, one of the primary challenges facing the development of such solutions is the identification of suitable data sources, i.e., data sets that could give a good contribution to the answer of user queries. In this paper, we discuss this problem and we present a feedback-based approach to incrementally identify new data sets for domain-specific linked data application.