Periodic variation model of pilotage volume in the Yangtze River based on neural network prediction

In order to study the problem of cyclic variation of pilotage volume, this paper proposes to study the long-cycle and short-cycle variation of the Yangtze River pilotage volume by the method of neural network prediction. Based on the dynamic analysis of the pilotage volume of the Yangtze River, this paper constructs the long-period and short-period variation models of the pilotage volume of the Yangtze River by means of the BP（Back Propagation） neural network empirical formula and repeated training method, and verifies the accuracy of the prediction of the BP neural network through the comparison of the regression fitting degree. The study on the variation rule of the Yangtze River pilotage volume shows that the long-period variation rule of the Yangtze River pilotage volume has a strong correlation with the I_BD（Baltic Dry Index）, while the short-period variation rule is affected by the domestic market and the Yangtze River shipping. The forecast study of pilotage volume provides reasonable data and decision support for the development planning and resource layout of pilotage organizations, so as to enable the port to develop safely in a high speed. The cycle change pattern will be applied to analyze the impact of shipping under the epidemic, which will provide scientific reference for precise epidemic prevention and control, and for pilotage standard and normative services.