摘要
聚类算法是数据挖掘中核心技术之一,而k-means算法在经典聚类算法中占有重要地位。针对k-means聚类算法的最佳聚类个数k不易获得,因而使得该聚类算法的应用受到限制,为此提出一种k值优化方法:通过给出大于最佳聚类数的可能聚类数,而得到优化的聚类个数。通过实例给予验证,其结果说明该方法合理有效。
clustering is the one of core technology in data mining. K-means algorithm is a very famous clustering algorithm in the classical clustering. The paper focus on the clustering number of k-means algorithm which is hard to be given and hinders the application. So the paper puts forward a novel k optimized method, which we can obtain the optimized number of clusters if we afford a maximum number of clusters. The test experiment has proved the method reasonable and right.
出处
《巢湖学院学报》
2007年第6期21-24,共4页
Journal of Chaohu University
关键词
聚类
K均值算法
聚类数优化
数据挖掘
Clustering
k-means algorithm
clustering number optimized
data mining
