| 一种对奇异值不敏感的ISOMAP |
| |
| 引用本文： | 魏莱,王守觉,徐菲菲.一种对奇异值不敏感的ISOMAP[J].计算机应用,2007,27(8):1959-1960. |
| |
| 作者姓名： | 魏莱 王守觉 徐菲菲 |
| |
| 作者单位： | 同济大学,计算机科学与技术系,上海201804 |
| |
| 基金项目： | 国家自然科学基金
,
高等学校博士学科点专项科研项目 |
| |
| 摘 要： | ISOMAP是一种经典的非线性降维方法，能够有效地发现高维非线性数据集的低维几何结构，但该算法对奇异值和噪声非常敏感。利用具有鲁棒性的主成分分析（Robust PCA）来探测奇异点，并对奇异点进行适当处理以降低ISOMAP对其的敏感程度。所提出的算法直观且易于理解，实验结果也证明它具有较好的鲁棒性，而且在奇异点较多的情况下仍能保持数据的整体结构。
|
| 关 键 词： | 流形学习 主成分分析 等度规映射 |
| 文章编号： | 1001-9081（2007）08-1959-02 |
| 收稿时间： | 2007-02-27 |
| 修稿时间： | 2007-02-272007-04-10 |
Robust ISOMAP insensitive to singular value |
| |
| WEI Lai,WANG Shou-jue,XU Fei-fei.Robust ISOMAP insensitive to singular value[J].journal of Computer Applications,2007,27(8):1959-1960. |
| |
| Authors: | WEI Lai WANG Shou-jue XU Fei-fei |
| |
| Affiliation: | Department of Computer Science and Technology, Tongji University, Shanghai, 201804 |
| |
| Abstract: | ISOAMP is a classical nonlinear dimensionality reduction algorithm. It is effective to discover the low-dimensional manifold in a high-dimensional data space. But the algorithm is very sensitive to the noises and singular value. Principal Component Analysis with robustness (Robust PCA) was used to detect singular points, and the singularity was also appropriately treated to reduce the ISOMAP's sensitivity to it. The proposed algorithm is intuitive and easy to understand, the results of the experiment prove that it is robust, and can maintain the overall structure of data with more singular points. |
| |
| Keywords: | manifold learning Principal Component Analysis （PCA） ISOMAP |
| 本文献已被 维普 万方数据 等数据库收录！ |
| 点击此处可从《计算机应用》浏览原始摘要信息 |
|
点击此处可从《计算机应用》下载全文 |
