|
| 摘要: |
| 聚类分析是模式识别的一个重要分支，本文以核心点和k-均值算法为基础，提出了一种基于参考点的快速k-均值算法。本算法以参考点作为第一个初始聚类中心，剩余初始聚类中心在核心点中选取，使得初始聚类中心能更好的反映模式样本集的几何特征，并且能减少迭代次数。 |
| 关键词: 参考点 密度 k-均值 |
| DOI： |
| 分类号: |
| 基金项目: |
|
| A Fast K-means Algorithm Based on Reference and Density |
|
LI You-ming
|
| Abstract: |
| Clustering analysis is an important branch of pattern recognition ，in?this?paper,?the?core-point?and?k-means?algorithm?as?the?foundation,?proposed?a?fast?k-means?algorithm?based?on?reference?and?density.?on?the?reference?point?as?the?first?initial?cluster?centers,?the?remaining?initial?clustering?center?in?the?core?points?are?selected,?the?initial?clustering?center?can?better?response?pattern?sample?set?of?geometric?features,?and?can?reduce?the?number?of?iterations. |
| Key words: reference -point density k- means |
