Taking All Positive Eigenvectors Is Suboptimal in Classical Multidimensional Scaling.
Jeffrey TsangRajesh PereiraPublished in: SIAM J. Optim. (2016)
Keyphrases
- multidimensional scaling
- principal component analysis
- low dimensional
- dimensionality reduction
- latent structure
- cluster analysis
- kernel pca
- euclidean distance
- covariance matrix
- vector space
- spectral clustering
- singular value decomposition
- pattern recognition
- high dimensional
- manifold learning
- latent semantic indexing
- data points