摘要： 提出一种分层聚类算法,该算法可识别任意形状、大小的类,在某电信企业的客户分析中取得了较好的结果。算法首先从不同的角度对电信客户进行聚类或分类,然后以这些类为基础,实行自底向上的层次聚类得到最终的聚类结果。算法执行效率高,适合大规模数据的聚类问题。

关键词: 数据挖掘, 聚类, 聚类算法

Abstract: This paper proposes a layered clustering algorithm that can identify clusters having non-spherical shapes and wide variances in size and has been successfully applied in the analysis of the user of a telecom corporation.The algorithm first clusters the user of telecom corporation from various aspects,and then based on these clusters on the hierarchical merging algorithm is performed to get the final results.The algorithm is efficient and can facilitate the clustering of a very large data set.

中图分类号: 

• TP311.1