Abstract: Clustering analysis is the core in data mining technology.By using the related visualization methods to visual clustering result，the distribution of data is presented in a directly and visually form in order to allow policy-maker to analysis data directly.I-Miner is an enterprise data mining tool，and it is used to clustering analysis or visual clustering result by kinds of methods.Through using the S language to develop software function，the K-Medoids，SOM，and SOM combining with K-Medoids algorithms are programmed.Particularly in the high dimensional data visualization，the star chart and the U-matrix by SOM are developed to deal with the defect of the software I-Miner about clustering and visual high dimensional data.

Key words: data mining, clustering analysis, visualization, self-organizing map

摘要： 聚类分析是数据挖掘中的核心技术，利用相关的可视化方法显示聚类结果，将数据分布以直观、形象的图形方式呈现给决策者，使得决策者可以直观地分析数据。I-Miner是一个企业级的数据挖掘工具，利用I-Miner软件进行聚类分析，并用多种方法将聚类结果可视化。通过S语言拓展软件功能，编程实现了K-Medoid算法、SOM算法、SOM与K-Medoids结合的聚类组合算法，尤其是在高维数据的可视化上，实现了星图法和SOM之U矩阵法，弥补软件中聚类和可视化模块较少的不足。

关键词: 数据挖掘, 聚类分析, 可视化, 自组织神经网络