SPARQLGen: generation of SPARQL from pseudo BGP

Abstract

SPARQL is the querying language and communication protocol for communicating with RDF data sources. SPARQL query requires knowledge of URIs of bound values in the triple patterns and ontological schema used by dataset. A person, even expert in SPARQL, nds hard to gure out URIs for bound values to be used in the query. This requirement brings a gap between end user and SPARQL query formation. In this work, we aim to facilitate semantic search over web of data by converting keywords into URIs, and present SPARQLGen. SPARQLGen provides an easy way of writing SPARQL query for a given query over Web of Data (RDF data). Through appropriate interface, semantic annotations of keywords are captured. We derive a Pseudo Basic Graph Pattern which is basically similar to SPARQL BGP except that it contains keywords rather than full resource URIs. Here, we propose heuristics that discover URIs for annotated keywords and build corresponding SPARQL query. SPARQLGen takes services of falcons, a semantic search engine. The Linked Open Data plays the major role in nding aliased URIs of an entity. The nal set of results contains a list of URIs of different data sources. SPARQLGen bridges the gap between end user and SPARQL query formation. The interface allows users to write user intended keywords instead of highly syntactic SPARQL query so that he/she needs not worry about the URIs of entities while writing their queries.