Mapping Natural Language Questions to SPARQL Queries for Job Search

A technique for enabling end users to explore semantically annotated data in job search domain, Sem-QAS is presented. It translates a natural language text query into SPARQL by semantically identifying distinct atomic filtering constraints and their semantic association present in the input query. Sem-QAS dynamically forms complex SPARQL queries by combining the triple patterns generated for atomic filtering constraints. The system maintains a high recall and precision by paying special attention to the processing of scope modifiers and association operators. The efficacy and correctness of Sem-QAS is evaluated using Mooney Job data set and queries collected from a real job search engine.