Computer Engineering and Applications ›› 2010, Vol. 46 ›› Issue (33): 212-214.DOI: 10.3778/j.issn.1002-8331.2010.33.060

• 图形、图像、模式识别 • Previous Articles     Next Articles

Adaptive semi-supervised fuzzy spectral clustering algorithm

DAI Yue-ming,GAO Qian   

  1. School of Information,Jiangnan University,Wuxi,Jiangsu 214122,China
  • Received:2009-03-24 Revised:2009-06-15 Online:2010-11-21 Published:2010-11-21
  • Contact: DAI Yue-ming

自适应半监督模糊谱聚类算法

戴月明,高 倩   

  1. 江南大学 信息工程学院,江苏 无锡 214122
  • 通讯作者: 戴月明

Abstract: Semi-supervised clustering employs a small amount of labeled data to aid unsupervised learning.The number of clusters is required in most existing algorithms.An improved semi-supervised clustering & adaptive algorithm is introduced to overcome the shortages of the most existing algoritms.The experiments show good results.

Key words: spectral clustering, semi-supervised, adaptive, Fuzzy Kernel C-Means(FKCM)

摘要: 半监督聚类利用少部分标签的数据辅助大量未标签的数据进行非监督的学习,从而提高聚类的性能。大部分的谱聚类算法都需事先确定聚类数目,利用半监督机器学习技术和自适应聚类算法,解决算法中存在的聚类数目需要事先确定、易陷入局部最优、收敛速度缓慢、对孤立点敏感等缺陷。实验证明该算法有很好的聚类效果。

关键词: 谱聚类, 半监督, 自适应, 模糊核C-均值(FKCM)

CLC Number: