A matrix-based methodology for determining a part family's learning rate.
William A. Young IIDale T. MaselRobert P. JuddPublished in: Comput. Ind. Eng. (2008)
Keyphrases
- learning rate
- learning algorithm
- weight vector
- convergence rate
- error function
- hidden layer
- adaptive learning rate
- rapid convergence
- training algorithm
- activation function
- multilayer neural networks
- convergence speed
- metaheuristic
- bp neural network algorithm
- natural gradient
- neural network
- delta bar delta
- convergence theorem
- feature space
- data mining