Research On Regional Logistics Demand Forecast For Heilongjiang Province Based On GA-SVR Model And Grey Theory

Under the background of the country's active promotion of the "the Belt and Road" strategy,Heilongjiang Province took the initiative to connect with the "the Belt and Road" strategy of the country with its superior location conditions,proposed the "Longjiang Silk Road Economic Belt" construction plan,and promoted the "The economic corridor of China,Mongolia and Russia " construction.As a key industry to ensure the normal operation of various trade activities in the society,the logistics industry promotes the healthy and s table development of the regional economy.Promoting the construction of "Longjiang Silk Road Economic Belt" is inseparable from the support of the logistics industry,and more needs a sound business logistics system as a guarantee.The forecast of regional logistics demand in Heilongjiang Province provides support for the rational allocation of regional logistics resources and the c onstruction of a scientific and efficient regional logistics system.It also provides guarantee for the improvement of the government's governance ability.It has practical guiding significance for reducing waste of resources,promoting the sustainable and healthy development of regional economy and guaranteeing the construction of "Longjiang Silk Road Economic Belt".Based on the relevant theory of regional logistics demand forecasting,combined with the actual development of regional logistics in Heilongjiang Province,this paper analyzes the relevant indicators of regional logistics demand in Heilongjiang Province,and uses the entropy weight grey correlation analysis method to screen regional logistics related indicators.Heilongjiang Province regional logistics forecasting indicator system.Four partial prediction models,partial least squares regression model,BP neural network model,GS-SVR model and GA-SVR model were established by partial least squares method,BP neural network and support vector mac hine method.Finally,the GA-SVR model was selected as Heilongjiang.Provincial regional logistics demand forecasting model.The grey forecasting method is used to predict the regional logistics demand related indicators of Heilongjiang Province,and the forecast results are substituted into the GA-SVR model to predict the logistics demand data of Heilongjiang Province in the next three years.Through the research and analysis of the regional logistics demand forecast of Heilongjiang Province,the main contents and conclusions of this paper are as follows:(1)Based on the literature research related to logistics forecasting,this study combines the research and analysis of the actual development of Heilongjiang Province,analyzes and summarizes the forecasting indicators of logistics demand in Heilongjiang Province and the main factors affecting the regional logistics demand in Heilongjiang Province,and selects the freight volume as the main factor.Heilongjiang Province logistics demand forecasting indicat ors,initially selected 20 logistics demand-related influencing factors indicators,using the entropy weight grey correlation analysis method to calculate the gray correlation degree of each impact indicator,and filter the relevant influencing factors.Finally,14 relevant impact indicators were selected to construct an indicator system for regional logistics demand forecasting in Heilongjiang Province.(2)Based on the research on the prediction method of regional logistics demand,this study selects three prediction methods: partial least squares,BP neural network and support vector machine.The model of grid search and genetic algorithm for support vector machine is selected.The establishment of relevant parameters optimization,combined with the estab lished Heilongjiang regional logistics demand forecasting index system,the partial least squares regression model,BP neural network regression model,GS-SVR model and GA-SVR model for regional logistics demand forecasting.By comparing and analyzing the GA-SVR model,the average relative error of the prediction results is 1.70%,and the prediction effect is the best.Finally,the regional logistics demand forecasting model of Heilongjiang Province based on GA-SVR model was selected.(3)The research analyzes the characteristics of the changes of the relevant impact indicators of regional logistics demand in Heilongjiang Province,selects the grey forecasting method as the appropriate forecasting method,and establishes the forecasting model of the regional logistics demand-related impact indicators in Heilongjiang Province,and the relevant impact indicators.Forecasts for the next three years.(4)Based on the established regional logistics demand forecasting index system and GA-SVR model of Heilongjiang Province,the research will substitute the forecast results of regional logistics related impact indicators predicted by the gray forecasting model into GA-SVR for the next 3 years.In the model,the logistics demand of Heilongjiang Province in the next thr ee years is predicted,providing scientific data support for government-related decision-making.