The p Operator: Discovering and Ranking Associations on the Semantic Web

In this paper, we introduce an approach that supports querying for Semantic Associations on the Semantic Web. Semantic Associations capture complex relationships between entities involving sequences of predicates, and sets of predicate sequences that interact in complex ways. Detecting such associations is at the heart of many research and analytical activities that are crucial to applications in national security and business intelligence. This in combination with the improving ability to identify entities in documents as part of automatic semantic annotation, gives a very powerful capability for semantic analysis of large amounts of heterogeneous content. The approach for supporting Semantic Associations discussed in this paper has four main facets. First, it generalizes these associations into three main classes based on their structural properties, allowing us to reason about them in a domain-independent manner. The second is the provision of an operator ρ for expressing queries about such associations. Third, it uses a graph data model for knowledge representation, allowing the semantic associations search techniques to be built upon the graph algorithms for paths, while integrating knowledge from the schema into the search process. The fourth facet is the use of a notion of context, which allows for restricting the search space and for context-driven ranking of results. Just as a Web search engine looks for relevant documents in the current Web, ρ can be seen as discovering and ranking complex relationships in the Semantic Web. In this paper, we demonstrate the need for supporting such complex semantic relationships. We also give a formal basis to the notion of Semantic Associations and give a brief discussion on our overall approach for discovering and ranking them