| 用于动态心电图波形分类改进的K近邻算法研究 |
| |
| 引用本文： | 苑静中.用于动态心电图波形分类改进的K近邻算法研究[J].天津师范大学学报(自然科学版),2008,28(3):60-63. |
| |
| 作者姓名： | 苑静中 |
| |
| 作者单位： | 天津理工大学,计算机科学与技术学院,天津,300191 |
| |
| 基金项目： | 天津市高等学校科技发展基金 |
| |
| 摘 要： | 针对动态心电图波形数据量大且具有明显个体差异性的特点,提出了一种改进的K近邻分类算法,用于动态心电图波形分类.该算法首先将实例间的度量改为曼哈顿距离（City Block Distance）,然后引入高斯核函数,将K近邻算法改进为非线性分类算法,以达到分类动态心电图波形的目的.实验结果表明,该算法在对动态心电图波形进行分类时,分类精度在90%以上.
|
| 关 键 词： | 分类 曼哈顿距离 K近邻算法 动态心电图 |
On improved K-nearest neighbor algorithm used for classification of waveforms of dynamic electrocardiogram |
| |
| YUAN Jingzhong.On improved K-nearest neighbor algorithm used for classification of waveforms of dynamic electrocardiogram[J].Journal of Tianjin Normal University(Natural Science Edition),2008,28(3):60-63. |
| |
| Authors: | YUAN Jingzhong |
| |
| Institution: | YUAN Jingzhong (School of Computer Science and Technology, Tianjin University of Technology, Tianjin 300191, China) |
| |
| Abstract: | An improved K-nearest neighbor algorithm used for classification of waveforms of dynamic electrocardiogram is proposed based on its features of abundant database and obvious individual difference.K-nearest neighbor algorithm is improved to nonlinear sorting algorithm by transforming the distance between samples into City Block Distance and bringing in Gaussian kernel function.The experiment results show that the classification precision of the algorithm is more than 90%. |
| |
| Keywords: | classification City Block Distance K-nearest neighbor algorithm dynamic electrocardiogram（Holter） |
| 本文献已被 维普 万方数据 等数据库收录！ |
|
