Progressive Semantic Query Answering

Ontology-based semantic query answering algorithms suffer from high computational complexity and become impractical in most cases that OWL is used as a framework for data access in the Semantic Web. For this reason, most semantic query answering systems prefer to lose some possible correct answers of user queries rather than being irresponsible. Here, we present a method that follows an alternative direction that we call progressive semantic query answering. The idea is to start giving the most relevant correct answers to the user query as soon as possible and continue by giving additional answers with decreasing relevance until you find all the correct answers. We first present a systematic analysis that formalises the notion of answer relevance and that of query answering sequences that we call strides, providing a formal framework for progressive semantic query answering. Then, we describe a practical algorithm performing sound and complete progressive query answering for the W3C’s OWL 2 QL Profile.