350 rub
Journal Neurocomputers №10 for 2011 г.
Article in number:
Neural network algorithms of behavior of respondents
Authors:
L. N. Yasnitsky, Z. I. Sichinava
Abstract:
First demonstrated the ability to create polygraph based on neural network technology. Shown that the use of neural network technology to build a polygraph can achieve higher accuracy of their findings compared with existing methods. Scopes of each of the three types of neural polygraph are devined. The application developed in the thesis algorithms in practice, the instrumental detection of deception reduces the need to attract skilled professionals, of polygraph, and also reduces the role of human factor-ray, allows users to use polygraph. All algorithms are based on neural networks of a perceptron type with sigmoidal activation functions.
Pages: 59-64
References
  1. Чечкин А. В. Сетевое моделирование проблемной области сложных систем - современный этап нейрокомпьютерных технологий // Нейрокомпьютеры: разработка, применение. 2011. № 2. С. 3-5.
  2. Петров А. М.Из записной книжки полиграфолога. Пермь: Издательский дом «Компаньон». 2003.
  3. Ясницкий Л.Н. Введение в искусственный интеллект. Изд. 3. М.: Издательский центр «Академия». 2010.