摘要
针对传统的K-均值聚类算法中随机选取初始聚类中心的缺陷,提出一种改进的K-均值聚类算法,利用自组织特征映射网络(SOFM)自动获得初始聚类中心。实验结果表明,改进的K-均值聚类算法能有效改善聚类性能,提高聚类的准确率。
In view of the shortcomings of traditional K-means algorithm in not being able to select the initial clustering center automatically, a new improved K-means clustering algorithm is proposed, which obtains the initial clustering center by using Self- Organizing Feature Map (SOFM) automatically. Experimental results demonstrate that the improved K-means algorithm can improve the clustering performance effectively and enhance the clustering accuracy.
出处
《科技导报》
CAS
CSCD
北大核心
2009年第10期61-63,共3页
Science & Technology Review
基金
安徽省高校省级自然科学基金项目(KJ2007B158)