| 基于划分采样的初始聚类中心算法 |
| |
| 引用本文： | 李玮,薛惠锋,詹海亮. 基于划分采样的初始聚类中心算法[J]. 太原理工大学学报, 2011, 42(4): 334-337 |
| |
| 作者姓名： | 李玮 薛惠锋 詹海亮 |
| |
| 作者单位： | 西北工业大学自动化学院,西安,710072 |
| |
| 基金项目： | 国家自然科学基金资助项目(70873079); 山西省自然科学基金资助项目(2009011021-1) |
| |
| 摘 要： | 针对大数据集的初始聚类中心选取问题,在基于密度的划分算法和适用于大规模数据集限定初值的采样算法基础上,提出了一种用于初始聚类中心的划分采样算法。对聚类子空间在每一维上进行均匀划分形成不同的数据区域,根据数据区域的数据点数的多少进行采样来提高采样的准确性。利用采样思想缩小了数据集的规模,保证了算法在时间上的优势。通过不同规模、不同形状的数据集对算法进行验证,实验结果表明,与其它初始聚类中心算法相比,该算法在准确率和时间上都具有一定的优势。
|
| 关 键 词： | 聚类中心初始化 密度估计法 限定初值算法 K-means算法 |
The Initial Cluster Center Algorithm Based on Segmenting and Sampling |
| |
| LI Wei,XUE Hui-feng,ZHAN Hai-liang. The Initial Cluster Center Algorithm Based on Segmenting and Sampling[J]. Journal of Taiyuan University of Technology, 2011, 42(4): 334-337 |
| |
| Authors: | LI Wei XUE Hui-feng ZHAN Hai-liang |
| |
| Affiliation: | (College of Automation,Northwestern Polytechnical University,Xi′an 710072,China) |
| |
| Abstract: | Based on the density segmenting algorithm and large scale data group segmenting algorithm,a new algorithm for initializing the cluster center was presented.The idea of segmenting base on density was combined with the idea of sampling and the new idea was presented.The accuracy of sampling was improved by averagely segmenting every dimension of the database.The speediness of the refined initial algorithm ensured the superiority of new algorithm on time.The experiment demonstrates that the new algorithm had s... |
| |
| Keywords: | cluster center initialization density estimation refining initial algorithm K-means |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
|
