Automatic Structuring of Radiology Free-Text Reports

摘要: A natural language processor was developed that automatically structures the important medical information (eg, existence, properties, location, and diagnostic interpretation of findings) contained in a radiology free-text document as formal model can be interpreted by computer program. The input to system is report from radiologic study. requires no reporting style changes on part radiologist. Statistical machine learning methods are used extensively throughout system. graphical user interface has been allows creation hand-tagged training examples. Various aspects difficult problem implementing an automated structured have addressed, relevant technology progressing well. Extensible Markup Language emerging preferred syntactic standard for representing distributing these reports within clinical environment. Early successes hold out hope that...