Computer Science and Information Technologies, Computer Science and Information Technologies 2012

Font Size: 
Optimization of training neural networks using genetic algorithm
Eugene A. Kaspransky, Olga A. Kazban, Olga S. Nurgayanova

Last modified: 2024-01-10

Abstract


One of actual problems of constructing neuralnetworks is creation of effective algorithm of thetraining, allowing for the minimum time to set up aneural network on recognition of the set of inputimages. To optimize the learning of neuralnetworks has been selected genetic algorithm. Theresult of work of the implemented algorithmshowed that training with a genetic algorithm ismuch faster than algorithm of return distribution.

Keywords


Optimization; training; neural networks; genetic algorithm;

References


1. Rutkovskay D., Pilinsky M., Rutkovsky L. "Neuralnetworks, genetic algorithms and indistinct systems",Moscow, 2006.

2. Haykin S. "Neural networks: full course", England,2006.

3. Duke V., Samoylenko A. "Data Mining: trainingcourse", St. Petersburg, 2001.

4. Gorban A., "Training of neural networks", Moscow,1990.

5. Kruglov V., Borisov V., "Artificial neural networks.Theory and practice", Moscow, 2001.

6. Kureichik V., "Genetic Algorithms. State. Problem.Perspectives", Proceedings of the Academy ofSciences. Tisa. 1999. Number 1.

7. Fedyaev O., Solomka Y., "Application ofgeneticalgorithms for training neural networks", Ukraine,2004.


Full Text: PDF