论文部分内容阅读
目的 探讨临床多变量类型资料 SAS聚类分析的方法。方法 采用 Jaccard系数进行样品间的相似性度量 ,构建非欧几里德距离矩阵 ,在 SAS数据步过程中完成上述计算 ,并生成分析数据集 ,通过调用 Cluster过程对资料进行聚类分析。结果 通过聚类分析 ,5 0例肝炎患者得到了合理归类 ,分类结果比较真实地反映了患者的临床特征。结论 利用 Jaccard系数度量样品间的相似性 ,构建距离矩阵的方法比较适用于临床多变量类型资料 ,操作简单 ,聚类效果也比较好。
Objective To explore the methods of clinical multivariate type SAS cluster analysis. Methods Jaccard coefficient was used to measure the similarity between samples, non-Euclidean distance matrix was constructed, the above calculations were completed during the SAS data step, and the analysis data set was generated. Cluster data analysis was performed by invoking the Cluster process. Results Through cluster analysis, 50 patients with hepatitis were reasonably classified. The classification results truly reflect the clinical characteristics of the patients. Conclusion Using the Jaccard coefficient to measure the similarity between samples, the method of constructing the distance matrix is more suitable for clinical multivariate type data. The operation is simple and the clustering effect is better.