MONTE CARLO SIMULATION OF SAMPLES WITH SPECIFIED PROPERTIES FOR TRAINING AND RESEARCH OF NEURAL NETWORKS OF DIRECT PROPAGATION

UDC 330: 45
DOI: 10.36871/2618-9976.2023.09.002

Authors

Olga V. Galanina,
Candidate of Economic Sciences, Saint Petersburg State Agrarian University, Saint Petersburg, Russia
Alexey M. Golubev,
ITMO National Research University, Saint Petersburg, Russia

Abstract

Today, big data is either unavailable or expensive. But it is necessary to investigate the behavior of neural networks of direct propagation right now. That is why it is necessary to model arrays of data (samples) with specified properties (for example, arrays of correlated quantities of any finite dimension) in order to investigate the behavior of neural networks of direct propagation of any architecture. It is possible to implement the task of forming such samples using the Monte Carlo method.

Keywords

Big data, Datasets, Direct propagation neural network, Perceptron, Monte Carlo method, Simulation modeling