作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程 ›› 2006, Vol. 32 ›› Issue (8): 105-106,123.

• 软件技术与数据库 • 上一篇    下一篇

基于一种可扩展函数族的聚类算法

樊建聪,梁永全,徐建国   

  1. 山东科技大学信息学院,青岛 266510
  • 出版日期:2006-04-20 发布日期:2006-04-20

Clustering Algorithms Based on an Extensible Function Family

FAN Jiancong, LIANG Yongquan, XU Jianguo   

  1. College of Information Science and Engineering, SUST, Qingdao 266510
  • Online:2006-04-20 Published:2006-04-20

摘要: 提出了可扩展函数族概念,将无关联的或者有依赖关系的多个函数,按照参数和自变量的特征集成一个函数族,可以通过增加或删除函数对函数族进行扩展。利用此函数族概念,设计了两种聚类算法,对其聚类性能作了分析。实验结果表明,用函数族方法进行聚类,在可伸缩性、增量数据处理和复杂数据类型处理等方面都表现出很好的性能,而且具有线性的时间复杂度。

关键词: 函数族;可扩展函数族;聚类

Abstract: An extensible function family is presented. It aggregates some functions. These functions have different parameters and variants. The relationship among the functions is dependent or independent. Adding or eliminating functions can extend a function family. With the conception,two clustering algorithms are designed. Experiments indicate that the algorithms have excellent performances in extensibility, incremental data processing and complex data type processing, and the time complexity is linear.

Key words: Function family; Extensible function family; Clustering