Weight vector jitter in the neural network with Hebb-s algorithm tuning

In this article the results of statistical analysis of neural network (NN) with Hebb-s algorithm tuning for weight vector jitter are represented. The expressions for correlation matrixes of output signals and output powers of neurons of the first, second and third neural network layer are obtained. It is shown, that the weight vector jitter leads to distortions of statistical characteristics of neural network. These distortions depends on number of neural network layer. The results of computer simulation, which confirming the conclusions of theoretical analysis are represented.