摘要
用Q型逐次信息群分对白银矿区42名矽肺患者和41名正常人头发样的元素谱Cr、Zn、Mg、Al、Cd进行无监督模式识别,获得分类清晰的谱系图,83个样本的判别正确率达98.8％。这一结果表明,元素谱的Q型逐次信息群分可望成为研究和预测矽肺病的一种新技术。
Unsupervised pattern recognition for the elements table, Cr,Zn,Mg,Al,Cd, in the hair species of the 42 silicotics and the 41 others in Baiyin diggings has been performed with Q-typestepwise informational cluster analysis and clearly classified dendrograms has been given out, in which, the accuracy is 98.8%. The results indicate that Q-type stepwise informational cluster analysis for the microelements table may be used as a new technique to research and predict anthrasilicosis.
出处
《计算机与应用化学》
CAS
CSCD
1995年第4期285-288,共4页
Computers and Applied Chemistry
基金
甘肃省自然科学基金资助项目
关键词
聚类分析
模式识别
矽肺
元素谱
人发
Cluster analysis, Information content, Pattern recognition, Anthrasilicosis, Elements table
