The modified algorithms of construction of neural network SOFM

The brief review of known neural network algorithms of construction of adaptive grids are presented here. Often used variants of functions of neurons adjacency are resulted. Dependence of quality of grid adaptation from function parameters has been researched on the basis of using Gaussian function about the distance between neurons in modified algorithm of network SOFM. According to computing experiments recommendations at the choice of function parameters of the adjacency at adaptation of flat grid on complex area by modified algorithm SOFM have been resulted here.