Abstract | ||
---|---|---|
The construction of similarity relationship a-mong data points plays a critical role in manifold learning. There exist two popular schemes, i.e., pairwise-distance based similarity and reconstruction coefficient based similar-ity. Existing works only have involved one scheme of them. These two schemes have different drawbacks. For pairwise-distance based similarity graph algorithms, they are sensitive to the noise and outliers. For reconstruction coefficient based similarity graph algorithms, they need sufficient sampled data and the neighborhood size is sensitive. This paper proposes a novel algorithm, called Local Neighborhood Embedding (LNE), which preserves pairwise-distance based similarity and reconstruction coefficient based similarity for finding the latent low dimensional structure of data. It has following three advantages: Firstly, it is insensitive to the choice of neighborhood size; Secondly, it is robust to the noise; Thirdly, It works well even in under-sampled case. Furthermore, the proposed objective function has a closed-form solution, which means it has a low computational complexity, and the experimental results illustrate that LNE has a competitive performance in dimensionality reduction. © 2013 ACADEMY PUBLISHER. |
Year | DOI | Venue |
---|---|---|
2013 | 10.4304/jsw.8.2.410-417 | JSW |
Keywords | Field | DocType |
dimension reduction,manifold learning,similarity graph,unsupervised learning | Graph algorithms,Embedding,Dimensionality reduction,Pattern recognition,Computer science,Outlier,Unsupervised learning,Artificial intelligence,Nonlinear dimension reduction,Nonlinear dimensionality reduction,Manifold,Machine learning | Journal |
Volume | Issue | Citations |
8 | 2 | 4 |
PageRank | References | Authors |
0.43 | 7 | 3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Liangli Zhen | 1 | 72 | 9.73 |
Xi Peng | 2 | 447 | 23.84 |
Dezhong Peng | 3 | 285 | 27.92 |