| 基于稀疏Parzen窗密度估计的快速自适应相似度聚类方法 |
| |
| 引用本文： | 钱鹏江,王士同,邓赵红.基于稀疏Parzen窗密度估计的快速自适应相似度聚类方法[J].自动化学报,2011,37(2):179-187. |
| |
| 作者姓名： | 钱鹏江 王士同 邓赵红 |
| |
| 作者单位： | 1.江南大学信息工程学院 无锡 214122 |
| |
| 基金项目： | 国家自然科学基金(60903100,60975027,60773206)资助~~ |
| |
| 摘 要： | 相似度聚类方法(Similarity-based clustering method,SCM)因其简单易实现和具有鲁棒性而广受关注.但由于内含相似度聚类算法(Similarity clustering algorithm,SCA)的高时间复杂度和凝聚型层次聚类(Agglomerative hierarchicalclu...
|
| 关 键 词： | 相似度聚类 密度估计 时间复杂度 图像分割 |
| 收稿时间： | 2010-5-11 |
| 修稿时间： | 2010-7-30 |
Fast Adaptive Similarity-based Clustering Using Sparse ParzenWindow Density Estimation |
| |
| QIAN Peng-Jiang,WANG Shi-Tong,DENG Zhao-Hong.Fast Adaptive Similarity-based Clustering Using Sparse ParzenWindow Density Estimation[J].Acta Automatica Sinica,2011,37(2):179-187. |
| |
| Authors: | QIAN Peng-Jiang WANG Shi-Tong DENG Zhao-Hong |
| |
| Affiliation: | 1.School of Information Technology, Jiangnan University, Wuxi 214122 ;2.School of Digital Media, Jiangnan University, Wuxi 214122 |
| |
| Abstract: | Similarity-based clustering method (SCM) has received much attention because it is robust and can be implemented simply and easily. However, because of its high time complexity of the embedded similarity clustering algorithm (SCA) and high space complexity of the embedded agglomerative hierarchical clustering (AHC ), SCM is impractical for large data sets. In this paper, the relationship is revealed between SCM and the kernel density estimation of samples, a novel fast adaptive similarity-based clustering method (FASCM) is accordingly proposed by adopting fast reduced set density estimator (FRSDE) and graph-based relaxed clustering (GRC). The distinctive advantages of FMSSC over MSSC exist in: 1) its asymptotic linear time complexity with the data size; 2) independent on artificial experience and its adaptability. Thus, FASCM is practical for large datasets. Its effectiveness has also been demonstrated in image segmentation examples. |
| |
| Keywords: | Similarity-based clustering density estimator time complexity image segmentation |
| 本文献已被 CNKI 等数据库收录！ |
| 点击此处可从《自动化学报》浏览原始摘要信息 |
|
点击此处可从《自动化学报》下载全文 |
