| 两类非线性降维流形学习算法的比较分析 |
| |
| 引用本文： | 王泽杰. 两类非线性降维流形学习算法的比较分析[J]. 上海工程技术大学学报, 2008, 22(1): 54-59. DOI: 10.3969/j.issn.1009-444X.2008.01.013 |
| |
| 作者姓名： | 王泽杰 |
| |
| 作者单位： | 上海工程技术大学,计算中心,上海,201620 |
| |
| 基金项目： | 上海市选拔培养优秀青年教师科研专项基金 |
| |
| 摘 要： | 流形学习(Manifold Learning)算法是近年来发展起来的非线性降维机器学习算法.目前的流形学习算法大体可以分为两类:全局的(如等度规映射)和局部的(如局部线性嵌套),它们有各自的优点和不足.以等度规映射(ISOMAP)和局部线性嵌套(LLE)为例,通过实验比较分析了这两类算法在参数选择、前提条件和执行效率上的特点,期望为不同应用提供参考.
|
| 关 键 词： | 流形学习 机器学习 非线性降维 等度规映射 局部线性嵌套 |
| 文章编号： | 1009-444X(2008)01-0054-06 |
| 修稿时间： | 2008-01-24 |
Comparison and Analysis of Two Categories of Manifold Learning Algorithms for Nonlinear Dimensionality Reduction |
| |
| WANG Zejie. Comparison and Analysis of Two Categories of Manifold Learning Algorithms for Nonlinear Dimensionality Reduction[J]. Journal of Shanghai University of Engineering Science, 2008, 22(1): 54-59. DOI: 10.3969/j.issn.1009-444X.2008.01.013 |
| |
| Authors: | WANG Zejie |
| |
| Affiliation: | WANG Zejie (Computer Center, Shanghai University of Engineering Science, Shanghai 201620, China) |
| |
| Abstract: | Manifold learning algorithms are nonlinear dimensionality reduction machine learning algorithms proposed in recent years.They fall broadly into two categories which have advantages and disadvantages respectively:global(such as ISOMAP)and local(such as LLE).Comparison and analysis of these two categories on parameter selection,premise and computational efficiency were given based on experimental results on ISOMAP and LLE.It is expected to give some insight into different applications. |
| |
| Keywords: | manifold learning machine learning nonlinear dimensionality reduction ISOMAP(isometric mapping) LLE(locally linear embedding) |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
| 点击此处可从《上海工程技术大学学报》浏览原始摘要信息 |
|
点击此处可从《上海工程技术大学学报》下载全文 |
