K-means聚类算法的研究综述

K-均值聚类算法（K-means）是基于划分的聚类算法中的典型算法，针对K-means算法初始聚类中心存在对K依赖的缺陷，提出一种新的选取K-means算法初始聚类中心的方法，该方法提高聚类结果的有效性和稳定性；还提出一种极值选择法，将最大距离法和最小距离法相结合，进一步提高初始聚类中心选择的准确性。

Summary of K-means Algorithm for Clustering

K-means clustering algorithm is partition-based typical algorithm for clustering , proposes a novel method for K-means clustering algo-rithm to select the initial cluster centers, for the purpose of drawback of K-dependent on the initial cluster centers in K-means clustering algorithm. This approach improves the effectiveness and stability of the clustering results. Proposes a selected method extreme that com-bines the K-means of maximum and minimum distance, to improve the accuracy of the initial cluster centers chosen.