聚类分析的新的启发式算法

A New Heuristic Algorithm for Clustering Analysis

对初始聚类点的选择提出一种有效的启发式算法.以内部聚类准则作为评价指标,实验结果表明,该算法明显好于 K-平均算法.

A heuristic algorithm is proposed for selecting initial seed values in the K-means algorithm.Re- sults show that the H-K algorithm outperforms the standard K-means algorithm when internal clustering cri- terion is computed as evaluation measures.