- Tytuł:
- Implementing evolutionary algorithm into training single-layer artificial neural network in classification task
- Autorzy:
- Płaczek, S.
- Powiązania:
- https://bibliotekanauki.pl/articles/95001.pdf
- Data publikacji:
- 2016
- Wydawca:
- Szkoła Główna Gospodarstwa Wiejskiego w Warszawie. Wydawnictwo Szkoły Głównej Gospodarstwa Wiejskiego w Warszawie
- Tematy:
-
genetic algorithm
evolutionary algorithm
selection process
mutation
recombination
replacement - Opis:
- The article proposes implementing a modified version of genetic algorithm in a neural network, what in literature is known as “evolutionary algorithm” or “evolutionary programming”. An Evolutionary Algorithm is a probabilistic algorithm that works in a set of weight variability of neurons and seeks the optimal value solution within a population of individuals, avoiding the local maximum. For chromosomes the real value variables and matrix structure are proposed to a single-layer neural network. Particular emphasis is put on mutation and crossover algorithms. What is also important in both genetic and evolutionary algorithms is the selection process. In the calculation example, the implementation of theoretical considerations to a classification task is demonstrated.
- Źródło:
-
Information Systems in Management; 2016, 5, 3; 377-388
2084-5537
2544-1728 - Pojawia się w:
- Information Systems in Management
- Dostawca treści:
- Biblioteka Nauki