A Refined Harmonic Lanczos Bidiagonalization Method and an Implicitly Restarted Algorithm for Computing the Smallest Singular Triplets of Large Matrices.
Zhongxiao JiaDatian NiuPublished in: SIAM J. Sci. Comput. (2010)
Keyphrases
- high accuracy
- cost function
- detection algorithm
- dynamic programming
- experimental evaluation
- clustering method
- detection method
- computational cost
- optimization algorithm
- synthetic and real images
- objective function
- preprocessing
- improved algorithm
- significant improvement
- estimation algorithm
- segmentation algorithm
- theoretical analysis
- classification method
- support vector machine svm
- computational complexity
- energy function
- input data
- tree structure
- matching algorithm
- sparse matrix
- high efficiency
- k means
- eigendecomposition
- computationally efficient
- em algorithm
- reconstruction method
- segmentation method
- decomposition method
- optimization method
- recognition algorithm
- classification algorithm
- probabilistic model
- np hard
- distance matrix
- coefficient matrix
- convergence rate
- selection algorithm
- maximum likelihood
- dimensionality reduction
- search space
- optimal solution
- similarity measure
- learning algorithm