| 摘要: |
| 许多经典的聚类算法,如平均链接,K-means,K-medoids,Clara,Clarans等,都是利用单一的聚类中心进行聚类.为克服单一聚类中心只能描述凸状聚类的缺陷,CURE,DBSCAN等算法使用多个代表点(或稠密点)表述任意形状的聚类结构,但仍难以聚类重叠和噪声数据.为此,提出一种基于多层聚类中心(称为核心集)的凝聚聚类算法(MulCA).该算法使用了多层核心集表述聚类结构,使得每一层数据集向其核心集凝聚.同时,上层的核心集自动成为下层的数据集.随着每层核心集规模按α比例迅速减少,控制了凝聚过程的迭代次数.此外,引入了基于随机采样计算ε-核心集(RBC)的技巧,将MulCA算法应用于大规模数据集.大量的数值实验充分验证了MulCA算法的有效性. |
| 关键词: 多层 核心集 凝聚 大规模 |
| DOI：10.3724/SP.J.1001.2013.04322 |
| 分类号: |
| 基金项目:国家自然科学基金(61103058, 61233011, 61272220) |
|
| Multilevel Core-Sets Based Aggregation Clustering Algorithm |
|
MA Ru-Ning1, WANG Xiu-Li1, DING Jun-Di2
|
|
1.College of Science, Nanjing University of Aeronautics and Astronautics, Nanjing 211100, China;2.School of Computer Science and Technology, Nanjing University of Science and Technology, Nanjing 210094, China
|
| Abstract: |
| Many classical clustering algorithms like Average-link, K-means, K-medoids, Clara, Clarans and so on are all based on a single cluster-center and are only apt to discover convex-structured clusters. Other methods, e.g., CURE and DBSCAN, use more than one point to represent a cluster and can find some well-separated clusters of arbitrary shape. However, they only consider the original scale of the input data; thus, they cannot depart over-lapped or noisy clusters. To this end, this paper is used to propose a multilevel core-set based agglomerative clustering algorithm (MulCA). The idea of MulCA is that the clustering structure is described by multi-level core set. Clustering process is achieved through procedure which the top of the core set automatically becomes the underlying data set. In addition, through the introduction of random sampling based ε-core set (RBC), MulCA algorithm is applied to large-scale data sets. A large number of numerical experiments fully verify the algorithm MulCA. |
| Key words: multilevel core-set aggregation large size |
