A sliding window solution for the on-line implementation of the Levenberg-Marquardt algorithm.
Fernando Morgado DiasAna AntunesJosé M. N. VieiraAlexandre Manuel MotaPublished in: Eng. Appl. Artif. Intell. (2006)
Keyphrases
- levenberg marquardt
- sliding window
- optimal solution
- fixed size
- least squares
- window size
- learning algorithm
- gauss newton
- convergence rate
- initial guess
- artificial neural networks
- neural network
- closed form
- boyer moore
- global optimum
- data streams
- nonlinear least squares
- compact data structure
- variable size
- training algorithm
- bundle adjustment
- back propagation
- objective function