计算机模式识别在十二指肠溃疡患者中微量元素谱上的应用

The Study of Computer Pattern Recognition Its Application in the Trace Elements of the Duodenum Ulcer

根据Mahalanobis距离判别法和W.Sammon提出的非线性映照法原理,用BASIC语言编制适用于TRS——80.APPLE和IBM——PC等微机的多元统计判别程序,对31例十二指肠溃疡患者和10例健康人的血清、胃液和胃粘膜中的铝、钼、铬、铜、铁、锶、锌、锰和钡等元素的含量高低进行分类判别。上述三类样本的分类准确率分别为90.2%、92.7%和100%。分类研究指出,十二指肠溃疡病变可能与人体中的钼、铝和铬有重要联系。

This paper brings forward a multivariate statistical classification decison method to determine the concentration of Al, Mo, Cr, Cu, Fe, Mn, Sr, Zn and Ba in 10 healthful human samples and 31 duodenum ulcer patient samples of serum, gastric juice and gastric mucous organism with a special program The program. On the basic of the Mahalanobis distans dicision and nonlinear mapping of W. Sammon, is made in BASIC language which is suitable for TRS-80, Apple and IBM-PC microcomputer. The right decision percentage of this three kinds of samples are 90.2%, 92.7% and 100% respectively. The research shows that the concentrations of Mo, Al and Cr in the body are greatly concerned in duodenum ulcer.