On the convergence of the LMS algorithm with a rank-deficient input autocorrelation matrix.
Desmond C. McLernonM. Mauricio LaraAldo G. Orozco-LugoPublished in: Signal Process. (2009)
Keyphrases
- preprocessing
- convergence rate
- np hard
- learning algorithm
- significant improvement
- computational complexity
- singular values
- segmentation algorithm
- input data
- computational cost
- optimal solution
- cost function
- k means
- high accuracy
- worst case
- dynamic programming
- matching algorithm
- recognition algorithm
- similarity matrix
- search space
- global convergence
- convergence theorem
- convergence proof
- update rule
- matrix representation
- learning management systems
- detection algorithm
- optimization algorithm
- simulated annealing