FUNDAMENTALS OF NEURAL NETWORK PREDICTION TECHNIQUES

The main theoretical material on the topic of introduction to neural networks is given. The most important, significant and widespread topologies of artificial neural networks (hereinafter referred to as NS) are consid- ered. The properties on which the NS is based are analyzed. Scopes of application, comparison of methodologies are defined. There are restrictions on the use of different models. Topologies: Kohonen, Grossberg, Hopfield, Hamming networks, the main type - perceptron are studied. The key architectures of these networks from both historical and practical point of view are considered.