论文部分内容阅读
一系列的概率性的计算机程序-使用从28000个病人取得的数据库-在根据心电图数据诊断心脏病时胜过了一群受过训练的心脏病学家。 这个系统从分析情况已经证实的病人的心电图正确地鉴别了病例的百分之八十八的疾病,而心脏病学家的分数只达百分之六十四。这些程序是由匹伯格博士(Dr.Hubert Pipberger)同华盛顿退伍军人管理局医院和佐治·华盛顿大学医学院有联系的临床工程和医学教授研究出来的。
A series of probabilistic computer programs - using a database of 28,000 patients - outperformed a group of trained cardiologists when diagnosing heart disease based on electrocardiogram data. The system correctly identifies 88% of the cases from the electrocardiogram of the patient whose condition has been confirmed, whereas the cardiologist scores only 64%. These procedures were developed by Dr. Hubert Pipberger, associate professor of clinical engineering and medicine at the Washington Veterans Administration Hospital and the George Washington University School of Medicine.