A method of analysis of the models of neural nets.

摘要: Taking into account Caianiello's work of 1961 a model neuron quite similar to his is proposed and studied. For this model, where temporal summation period refractoriness are assumed, mathematical approach simulation on computer were realized. Particular types nets used, namely: with topological structures, fully random nets. The difference between the two that first type has two-dimensional square structure depends rules formation connection neurons, while second realized by means probability distribution function governing net. These neural analysed method which permits obtain various parameters characterize their behaviour in time space terms trajectory system. Many experiments also reported; statistical analyses, made them, show great importance influence networks. In last part an interesting case reported, reaction net disturbance shows kind adaptation takes place, although stays unchanged.