Ranking Query Results based on the Query Semantics and in the Presence of Linkage and Evolution

Keyword queries offer a convenient alternative to traditional SQL in querying databases with large, often unknown, schemas and instances. The challenge in answering such queries is to discover their intended semantics, construct the different interpretations in the native query language and rank the results. We are going to present a framework for understanding and ranking the possible interpretations that, in contrast to existing keyword-based techniques, is not using any a-priori knowledge on the instance data. Once an interpretation is found, we will present how it can be answered on a database in the presence of linkage information or of semantic evolution relationships among the data. The method considers the different alternatives and ranks the data results according to the likelihood that they are actually representing the data that the user was actually looking.