摘要
针对传统聚类方法存在的不足，提出了一种基于自适应膨胀因子的聚类新方法（CAIF）。相对于现有的一些聚类方法，CAIF方法不需要用户确定类数k值，克服了聚类结果对k值的敏感性；与数据的输入顺序无关，能够进行增量聚类。同时，CAIF方法还能有效地发现孤立点。最后通过实验验证了该方法的有效性。
In the face of the deficiencies in the traditional clustering methods, this paper puts forward a new clustering based on adaptive inflation factor(CAIF) approach which is different from some existent clustering methods. This approach does not depend on the number of class sorts that has big influence on clustering result. CAIF, which is independent of data input order, can be used in incremental clustering. At the same time, CAIF can find isolated points efficiently. The experiments sufficiently prove the validity of this approach.
出处
《计算机工程》
CAS
CSCD
北大核心
2003年第9期100-102,共3页
Computer Engineering
关键词
自适应
膨胀因子
增量聚类
Adaptive
Inflation factor
Incremental clustering
