Article Preview
TopThe goal of Linked Data query processing is an online execution of declarative queries over the SW, by relying only on the Linked Data principles (Bizer et al., 2009). Aiming to unleash the potential of SW, a number of general methods for querying Linked Data have been developed.
Early querying methods like Sindice (Oren et al., 2008) and Falcons (Cheng & Qu, 2009) are based on keywords. Since SPARQL (Harris & Seaborne, 2010) can be used to express queries across diverse data sources, it becomes the foundation of many recent researches about Linked Data querying. The devices in (Umbrich et al., 2011; Wagner et al., 2012) use index-based source selection and provide source ranking, while the techniques mentioned in (Hartig, 2013; Miranker et al., 2012) are traversal-based query execution approaches without source ranking. Moreover, some domain-oriented or datasource-oriented methods are presented to further pursue the accuracy and efficiency. NAGA (Mahdisoltani et al., 2014) provides best-effort heuristics to return and rank the relevant RDF triples from YAGO (Hoffart et al., 2013). By using SPARQL and SPIN, Lo Bue and Machi (Lo Bue & Machi, 2015) study on integrating and querying over tourism domain datasets via interlinking techniques. MEQLD (Tran & Nguyen, 2015) investigates to improve the mapping extension of lexical entities into DBpedia’s components for creating query in SPARQL. Besides, other current works focus on mapping visual method to SPARQL (Haag et al., 2015), optimizing query results over several heterogeneous Linked Data sources (Taelman, 2016), as well as automatically generating SPARQL query (Alec et al., 2016).