Comparison To Other Methods
Semidefinite embedding is much better in revealing the underlying dimension of the data compared to LLE and Laplacian eigenmaps. It also guarantees that the nearest neighbors in the embedding is the same as the original nearest neighbor for each point while the other two methods do not. On the other hand, semidefinite embedding is much slower and harder to scale to large data.
Semidefinite embedding outperforms Isomap when the manifold is not a convex subset of the Euclidean space.
Read more about this topic: Semidefinite Embedding
Famous quotes containing the words comparison and/or methods:
“Intolerance respecting other people’s religion is toleration itself in comparison with intolerance respecting other people’s art.”
—Wallace Stevens (1879–1955)
“If men got pregnant, there would be safe, reliable methods of birth control. They’d be inexpensive, too.”
—Anna Quindlen (b. 1952)