Implementation and performance evaluation of the agent-based algorithm for ANN training.
Ireneusz CzarnowskiPiotr JedrzejowiczPublished in: Int. J. Knowl. Based Intell. Eng. Syst. (2010)
Keyphrases
- learning algorithm
- artificial neural networks
- training algorithm
- dynamic programming
- detection algorithm
- theoretical analysis
- k means
- computational complexity
- worst case
- np hard
- high accuracy
- parallel implementation
- optimization algorithm
- linear programming
- experimental evaluation
- significant improvement
- preprocessing
- computational cost
- data sets
- feed forward neural networks
- search space
- computationally efficient
- objective function
- training process
- matching algorithm
- hardware implementation
- improved algorithm
- levenberg marquardt
- cost function
- training phase
- feature selection
- recognition algorithm
- clustering method
- multi agent
- segmentation algorithm
- training set
- simulated annealing
- support vector machine