论文部分内容阅读
作者分别用逐步回归法、临床经验法和逐步回归结合临床经验法三种方法选取了30、31和31个项目,然后采用逐步判别法和Fisher判别法两种统计方法建立起4个计算机诊断系统。经诊断303例回代样本和70例新样本后发现,逐步回归法作为项目的选取方法最好;逐步回归结合临床经验法次之;临床经验法最差;而两种统计方法均可作为建立计算机诊断系统的有效方法。最后作者认为,只考虑临床因素而不考虑数学因素的项目选取方法不够理想。
The author selected 30, 31, and 31 items using three methods: stepwise regression, clinical experience, stepwise regression, and clinical experience, and then established four computer diagnostic systems using stepwise and Fisher discriminant methods. . After the diagnosis of 303 cases of subsequent samples and 70 new samples, it was found that the stepwise regression method was the best method for selecting items; stepwise regression combined with clinical experience method was the second; clinical experience method was the worst; and both statistical methods could be used as establishment An effective method of computer diagnostic systems. Finally, the author believes that the project selection method that only considers clinical factors without considering the mathematical factors is not ideal.