Tóm tắt:

This article presents the method for training a deep artificial neural network (DANN) for regression analysis; this method is based on the generalized delta rule. To illustrate the rule, a DANN with two hidden layers is used. The model’s construction is described in the form of mathematical equations. Subsequently, a DANN program is written in Visual C# .NET. This program is tested with the task of estimating the shear strength of soil samples.