Towards a canonical representation for machine understanding of natural language in radiology reports. Bashyam, Vijayaraghavan. 原文请求 摘    要： A large amount of clinical data is recorded in health-care centers in the United States as a result of routine clinical care. This collection of data captures snapshots of disease processes across a large number of patients over time, presenting the possibility for discovering new scientific knowledge. However, the utility of this information is largely unfulfilled as the majority of today's clinical data is not directly amenable to scientific analysis, primarily due to the largely unstructured representation of the medical record contents that are not directly computer understandable. This dissertation investigates methods to transform anatomical knowledge found in free-text radiology reports to a representation more suitable for computational analysis. 分 类 号： 第二语种摘要： 正文语种： ENG 授予单位： University of California, Los Angeles. 学科专业： 导师姓名： Leazer, Gregory H., 授予学位： Ph.D. 授予年份： 2008 页    码： 140 p. 馆 藏 号： AAI3356507