An Improved K-means Algorithm Based on Semi-Supervised and Leader

摘要：

To solve the problem of sensitivity to the initial center for K-means, we combine the semi-supervised learning with Leader method and propose an improved algorithm (I.e.SSLK).In this paper.we choose several optimal leaders from a small amout of labeled dafa.and the Leader method can keep the distribution characteristics of the data object Experiments on the UCI database show that the combined method outperforms the original K-means clustering.lt obtains the increase of U.6％,9.95％ and 0.93％ in precision, compared with the classical K-means on the dataset of iris.image and balance-scale respectively.At the same time.the proposed method is uesd in the UCI KDD database .and the experimental results also show the proposed method outperforms the K-means.

关键词： Clustering K-means algorithm Semi_Supervised learning Leader method

作者: Zhang Yan-Ping Zhang Juan Zhang Li-Na Chu Wei-Cui

作者单位: School of Computer Science and Technology, Anhui University, Hefei, 230039,China

会议类型: 国际会议

会议名称: 2011 3rd International Conference on Computer and Automation Engineering(ICCAE 2011)(2011年第三届IEEE计算机与自动化工程国际会议)

会议地点: 重庆

会议语种:英文

页码: 142-145

在线出版日期: 2011-01-21（万方平台首次上网日期，不代表论文的发表时间）