Incremental and Directed Rule-Based Inference on RDFS

The Semantic Web contributes to the elicitation of knowledge from data, and leverages implicit knowledge through reasoning algorithms. The dynamic aspect of the Web pushes actual batch reasoning solutions, providing the best scalability so far, to upgrade towards incremental reasoning. This paradigm enables reasoners to handle new data as they arrive. In this paper we introduce Slider-p, an efficient incremental reasoner. It is designed to handle streaming expanding data with a growing background knowledge base. Directed reasoning implemented in Slider-p allows to influence the order of inferred triples. This feature, novel in the state of the art at the best of our knowledge, enables the adaptation of Slider-p's behavior to answer at best queries as the reasoning process is not over. It natively supports rho df and RDFS, and its architecture allows to extend it to more complex fragments with a minimal effort. Our experimentations show that it is able to influence the order of the inferred triples, prioritizing the inference of selected kinds of triples.