A modification on minimum restarting method in the Arnoldi algorithm for computing the eigenvalues of a nonsymmetric matrix.
Hashem Saberi NajafiH. GhazviniPublished in: Appl. Math. Comput. (2006)
Keyphrases
- high accuracy
- cost function
- dynamic programming
- detection algorithm
- experimental evaluation
- improved algorithm
- detection method
- optimization algorithm
- k means
- preprocessing
- computationally efficient
- significant improvement
- segmentation algorithm
- computational complexity
- objective function
- clustering method
- computational cost
- synthetic and real images
- theoretical analysis
- classification algorithm
- recognition algorithm
- estimation algorithm
- distance matrix
- similarity measure
- learning algorithm
- input data
- tree structure
- matching algorithm
- classification method
- optimization method
- convergence rate
- support vector machine svm
- correlation matrix
- energy function
- segmentation method
- principal components
- qr decomposition
- transformation matrix
- decomposition method
- probabilistic model
- covariance matrix
- np hard
- pairwise
- lower bound
- particle swarm optimization
- coefficient matrix
- singular value decomposition
- support vector machine
- null space
- simulated annealing
- maximum likelihood
- gram matrix
- eigenvalues and eigenvectors
- graph cuts