Patch Alignment for Graph Embedding

摘要: Dozens of manifold learning-based dimensionality reduction algorithms have been proposed in the literature. The most representative ones are locally linear embedding (LLE) [65], ISOMAP [76], Laplacian eigenmaps (LE) [4], Hessian (HLLE) [20], and local tangent space alignment (LTSA) [102]. LLE uses coefficients, which reconstruct a given example by its neighbors, to represent geometry, then seeks low-dimensional embedding, these coefficients still suitable for reconstruction. preserves global geodesic distances all pairs examples.