Automated query transformation for searching semantically rich structured collections

The availability of semantically rich structured resources makes the search of such resources on the web even more interesting, where the formation of a meaningful structured query is now possible for better retrieval. Incorporation of concepts like role, category, topic, class, attributes, etc. to indicate the search target and constraints in a structured query enable better definition of information needs. However, the varieties of usages in application contexts create heterogeneity of structures definition and presentation, thus making it hard of using this information when formulating structured query for a search process. Hence, we are motivated to automate transformation of unstructured queries (which are more common for web user), to structured queries (which support better specification of information needs). This paper first explores two important issues arisen in the process of query transformation, i.e. multiple structures (with respect to the needs of handling multiple structured queries type) and multiple semantics (with respect to the needs of handling different collections of structured resources). Along with these issues, we define research problems, followed by an approach of mediating multiple structures and multiple semantics to support the query transformation from unstructured to structured form.