Decision makers, such as natural resource managers charged with managing our national forests or clinicians involved in patient care, need to find relevant information quickly and reliably. In the best case, they need to access and take benefit from up-to-date information on similar issues, decision, and situations. This exploratory project is focused on providing quick access to documents in a domain-specific setting.
Professionals write, index, and search for documents using specialized terminology (that often includes the use of normal English words) that may be unfamiliar to users with less domain expertise. Yet both experts and non-experts need to be able to find and use the information contained in these documents. This approach will exploit the various well developed, widely used controlled vocabularies and classification schemes to assist with indexing domain-specific documents.
Given the difficulty of information retrieval in general, the project is high-risk, but may succeed because the work occurs in a domain-specific setting with well-developed terminology. The intellectual merit of this proposal stems from the use of a novel thesaurus model, Metadata++, with path-based representation of term senses to assist with indexing. The broader impact of this project stems from the ubiquity of the problem of helping decision makers find information and from the extensive collaborations with the US Forest Service, the non-profit Nature Conservancy, and Denmark's Royal School of Library and Information Science.
