Industrial Automatic Control Systems and Controllers Annotation << Back Prediction Suburban Transport Flows Based on a Neural Network N.G. Kuftinova, A.V. Ostroukh, N.E. Surkova, K.A. Barinov The article proposes a prediction method based on neural networks to solve the problem of traffi c prediction of network congestion. Of the existing forecasting methods, the Box-Jenkins method (ARIMA, ARMA), which is based on one factor, is the most applicable for determining the throughput on any fragments of the transport network. The system analysis of external transport links focuses on the fact that the traffic flow is influenced by one of the main several factors as time. The presented neural network architecture is capable of displaying the dynamics and complexity of the traffic flow. A reliable and accurate system for predicting the congestion of the transport network is essential for the reliable functioning of the intelligent transport infrastructure. From a mathematical point of view, the problem can be formulated as the problem of predicting some observable value, for example, the speed or flow rate, at a given node of the transport network for a selected forecast horizon. Keywords: neural network; forecasting methods; network congestion traffic; traffic flow; neural network structure. DOI: 10.25791/asu.11.2020.1235 Pp. 40-45.