摘要
针对k-means算法在聚类过程中受初始聚类中心影响很大的问题,本文提出了一种优化初始聚类中心的方法。此方法通过计算聚类中心与其他各个点之间的距离,依次找到最佳的一组初始聚类中心组合。实验表明改进后的k-means算法提高了检测率,降低了误检率,产生了质量较高的聚类结果。
In allusion to the problem of k-means algorithm that is greatly affected by the initial clustering center, a new method is proposed to optimize the initial clustering center. The method calculating the distance between the clustering center and other points will find the best clustering center combination. Experiments on the web-log show that the improved k-means algorithm can improve the detection rate, reduce error rate, and produce a high clustering result.
出处
《微计算机信息》
2012年第10期431-432,468,共3页
Control & Automation
