A new method to stabilize fast RLS algorithms based on a first-order model of the propagation of numerical errors.
Ahmed BenallalAndré GilloirePublished in: ICASSP (1988)
Keyphrases
- theoretical analysis
- statistical model
- objective function
- mathematical model
- probabilistic model
- similarity measure
- linear models
- hybrid method
- sensitivity analysis
- computational efficiency
- classification algorithm
- computational cost
- modeling method
- evaluation method
- computationally efficient
- prediction model
- reconstruction method
- linear model
- numerical methods
- energy function
- cost function
- significant improvement
- autoregressive
- statistical methods
- classification method
- prior knowledge
- em algorithm
- recursive least squares
- model free
- gaussian distribution
- data structure
- learned models
- bp neural network
- propagation model
- optimization method
- support vector machine
- higher order
- markov random field
- input data
- parameter estimation
- bayesian networks
- segmentation method
- learning algorithm
- dynamic programming
- energy functional
- test data
- recommendation algorithms
- artificial neural networks
- least squares
- regularized least squares
- convergence rate