| 基于聚类数和初始值的K-means算法改进研究 |
| |
| 引用本文： | 屈新怀,高万里,丁必荣,李朕.基于聚类数和初始值的K-means算法改进研究[J].组合机床与自动化加工技术,2011(4). |
| |
| 作者姓名： | 屈新怀 高万里 丁必荣 李朕 |
| |
| 作者单位： | 合肥工业大学机械与汽车工程学院,合肥,230009 |
| |
| 摘 要： | 原始的K-means算法,随机生成初始质心,事先给定聚类数k,在该前提下进行聚类,大大降低了聚类的效果.文章是对原始K-means算法的改进,提出了一种基于密度选取初始质心和采取遗传算法优化聚类数k的算法.该算法在一定程度上解决了初始质心和聚类数k对聚类精度和效率的影响,提高了聚类的准确率.最后文章通过实验证明了改进算法的有效性.
|
| 关 键 词： | K-means算法 初始质心 聚类数k |
The K-means Algorithm Improvement Base on the Number of Clustering k and Initial Centroid |
| |
| QU Xin-huai,GAO Wang-li,DING Bi-rong,LI Zhen.The K-means Algorithm Improvement Base on the Number of Clustering k and Initial Centroid[J].Modular Machine Tool & Automatic Manufacturing Technique,2011(4). |
| |
| Authors: | QU Xin-huai GAO Wang-li DING Bi-rong LI Zhen |
| |
| Affiliation: | QU Xin-huai,GAO Wang-li,DIMG Bi-rong,LI Zhen(School of Machinery and Automobile Engineering,Hefei University of Technology,Hefei 230009,China) |
| |
| Abstract: | In the original K-means algorithm,clustering can be done with the initial centroid generated randomly and the number of clustering k given in advance,and it reduces the effect of clustering greatly.In this paper,we improve the original K-means algorithm,and propose a new algorithm that the initial centroid is chose by the density and the number of clustering k is optimized by genetic algorithm.To a certain degree,this algorithm minimize the effects of initial centroid and the number of clustering k on the c... |
| |
| Keywords: | K-means algorithm initial centroid clustering k |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
|
