Research on international trade logistics prediction based on back propagation neural network

Abstract

The development of international trade depends to a large extent on the progress of international logistics. However, international logistics cannot exist independently of international trade. Without goods provided by international trade, international logistics loses its foundation. Therefore, in order to accurately assess the demand for international logistics, it is necessary to have a detailed understanding of the development of international trade and to predict its future trends accordingly. In this work, we utilize backpropagation neural networks to predict trends and needs in international trade logistics. Specifically, we build a multi-layer perceptron model, which selects a variety of input variables such as goods circulation, economic indicators, trade policies, and seasonal factors. By training this model, it is possible to effectively learn and capture the complex relationships that affect international trade logistics from historical data. In the experimental analysis, the model has been repeatedly trained and adjusted, and finally demonstrated high accuracy and reliability.