350 rub
Journal Neurocomputers №10 for 2009 г.
Article in number:
Factors of influence on stability of processing regime of automatic perceptron eduation
Keywords: perceptron eduation effect of stability
Authors:
N. A. Ivanova
Abstract:
The article shows that process of perceptron eduation loses it-s stability if the number of parameters is increased and their quality is lowered. The automat of perceptron education exercise unstable fluctuations. Increase of number of examples to be taught with lead to reverse effect of stability increase and escape from the zone of unstable fluctuations. The perceptrone was supposed to be taught in regime of unstability whil rendering the results of continious search of the automat of education up to the multidimensional surface of optimization.
Pages: 40-45
References
  1. Хайкин С. Нейронные сети: полный курс, 2-е издание: Пер. с англ. М.: Издательский дом «Вильямс». 2006. 1104 с.
  2. Уосермен Ф. Нейрокомпьютерная техника. М.: Мир. 1992.
  3. Волчихин В.И., Иванов А.И., Фунтиков В.А. Быстрые алгоритмы обучения нейросетевых механизмов биометрико-криптографической защиты информации. Пенза; Изд-во Пензенского государственного университета. 2005. 273 с.