| 摘要: |
| 利用谱降维方法可以归结为求解带权图Laplacican矩阵L(G)的特征值λn-t对应的特征向量这一理论,通过代数方法估计λn-t的下界,并讨论非带权图的情况下λn-t的下界. |
| 关键词: 带权图 Laplacian矩阵 特征值 降维理论 |
| DOI： |
| 投稿时间：2010-04-19 |
| 基金项目:国家自然科学基金项目(60903131)资助 |
|
| Eigenvalue Estimation in Dimensionality Reduction Theory |
|
|
|
| Abstract: |
| The application of spectral dimension reduction method can be attributed to solve the eigenvectors corresponding to the eigenvalues of weighted graph Laplacican matrix L(G).The lower bound of λn-t is given by using algebra method,and discussed the case of non-weighted graph. |
| Key words: weighted graph Laplacian matrix eigenvalue dimensionality reduction theory |
