FIELD: computer equipment.
SUBSTANCE: invention relates to computer engineering and can be used to interpret the operation of models of artificial neural networks. Method comprises steps of obtaining at least one artificial neural network pre-trained on a set of objects; generating for each layer of the trained neural network at least one decision tree, wherein the decision tree receives as an input activation data of the corresponding layer obtained when passing through the neural network of the object, from the available data set; by means of decision trees prediction is made with the same response, which is provided by trained artificial neural network on this object; then obtaining for each object an ordered sequence of leaf numbers formed at the previous step of decision trees; further, a set of rules is formed, which predicts a sequence of leaf numbers on the object.
EFFECT: high quality and accuracy of interpreting the operation of the artificial neural network.
12 cl, 8 dwg
