With the explosion of medical information accessible via the Internet, there is a growing need for development of better access to the online medical literature databases through user-friendly systems and interface. The proliferation of online information and the diversity of interfaces to data collections has led to a medical information gap between medical researchers and the accessibility of medical literature databases. Users who need access to such information must visit a variety of sources, which can be both excessively time consuming and potentially dangerous if the information is needed for treatment decisions. In addition, information generated by using existing search engines is often too general or inaccurate. Particularly frustrating is that simple queries can result in an excessive number of documents retrieved - too many to search through to determine which are and which are not relevant. The goal of this research is to extend a bridge across the medical information gap by creating easy-to-use interfaces to medical literature databases based on UMLS-enhanced Semantic Parsing and Personalized Medical Agent (PMA): (1) UMLS-enhanced Semantic Parsing: Our first goal will be to combine noun phrasing and co-occurrence analysis techniques recently developed by The University of Arizona Artificial Intelligence Lab (AI Lab) for the NSF-funded Illinois Digital Library Initiative (DLI) project with existing components found in the Unified Medical Language System (UMLS) developed by NLM. (2) Personalized Medical Agent: The second goal will be to develop a dynamic, intelligent medical agent interface to assist searchers in effortlessly locating documents and summarizing topics in the documents. The interface is particularly suited for busy physicians.
| Hripcsak, George; Rothschild, Adam S (2005) Agreement, the f-measure, and reliability in information retrieval. J Am Med Inform Assoc 12:296-8 |
