聚类分析法在癌胚抗原数据挖掘分析中的应用研究

Clustering analysis and its application in data mining of carcinoembryonic antigen

目的探讨聚类分析方法在癌胚抗原数据的应用,以获取有价值的诊断信息。方法回顾分析自2008年4月至2013年4月癌胚抗原数据,采用SPSS Clementine数据挖掘软件的K-means模型,对CEA相关的性别、年龄等特征用聚类分析法进行挖掘,分析癌胚抗原及关联特征的临床意义。结果K-means算法对12532行癌胚抗原数据的挖掘结果分5组,其中3组有医学意义,特别是平均年龄为88.541岁的高龄男性老年患者罹患癌症的概率,较老年女性及低龄老年男性更高,临床上应加大对该部分人群的体检力度,争取早发现早治疗。结论通过对血清癌胚抗原结果进行数据挖掘,发现聚类结果符合医学意义,可实现精准的预防及治疗。

Objective To study the application of clustering analysis on carcinoembryonic antigen data in order to obtain valuable diagnostic information. Methods Carcinoembryonic antigen data from 2008 April to 2013 April in our hospital were collected to analyze the clinical significance of the tumor antigen and related characteristics （sex, age and other characteristics） by the analysis based on SPSS Clementine K-means. Results The carcinoembryonic antigen results of K-means algorithm for the data mining of 12532 were divided into 5 groups, including 3 groups with medical significance. Elderly male patients with an average age of 88. 541 year old were with higher cancer risk than the older women and younger elderly men. Clinical research should increase the intensity of physical examination on the part of the crowd, and strive for early detection and early treatment. Conclusions Through data mining on serum carcinoembryonic antigen results, we know that we may achieve precise prevention and treatment since the clustering results are consistent with the medical significance.