The identification of missing information resources through the query difference operator

Abstract

In this paper we consider the processing of queries posed over multiple information resources that advertise their contents in terms of globally available, domain-specific ontologies. We describe a technique to identify the exact portion of a user's query that may not be answered by the set of available information agents. This is achieved by reasoning over the advertisements of the agents relative to the user's query. Our technique is based on the realization that the set difference of the queries q 1 and q 2 may be computed as a syntactic manipulation of the expressions q 1 and q 2 for a well defined subset of the relational algebra over a restricted class of relational schemas. That is to say, one may, without materializing data, take the expressions for q 1 and q 2 , apply the query difference formula to yield q 3 , and be guaranteed that q 3 is logically equivalent to q 1 , q 2 . With this Query Difference operator defined, the ability to compute query intersection, subsumption and equivalence follow. These claims are formally defined and proven and an example from an online movie guide domain is provided.In addition to the identification of missing resource agents, we anticipate a number of other applications of the Query Difference operator. This includes, but is not limited to, limiting the generality of dynamically constructed user queries, efficient query planning, and monitoring and controlling access to sensitive information.