LATENT SEMANTIC INDEXING IN SUPPORT OF DATA RETRIEVAL

Mayo Clinic Coll Of Medicine, Rochester, Rochester MN

Investigators

Linked publications & trials

Abstract

We propose to extend the successful work we have achieved with statistically based indexing and retrieval systems, by incorporating semantic structures which accommodate the modifying attributes of clinical conCepts. Patient data is rarely limited to a single axis of meaning or detail, and retrieval for application in quality improvement, decision support, or epidemiologic research, demands Consistent information struCture. This proposal will invoke the knowledge and tool suites of the UMLS Specialist Lexicon, the SGML markup and recognition capabilities of the TextMachine application, extensions to our locally developed CliniCal Query Language, and layer these enhancements upon our core techniques for statistically based indexing and retrieval of patient data. We commit these activities to remain compliant with emerging standards for medical concept representation arising from the Canon efforts and the standardization processes at ANSI-HISPP, CEN TC251 and the CPRI initiatives.