Application of genetic algorithm for determining effective configuration of artificial neural networks

Described the application of genetic algorithm for determination of effective configuration of artificial neural networks, components of artificial neural networks (nodes - neurons and their interconnections) are determined. Possible options for configuring neural networks (number of inputs, number of outputs, base architecture, number of layers and neurons in them) are given. Step by step description of the genetic algorithm. The algorithm of application of the genetic algorithm for the determination of the configuration of artificial neural networks is presented and its efficiency is substantiated. The practical aspects of using this approach when using input from different sources are given.