Research On Shanghai Logistics Demand Forecasting Based On Grey Theory And KPCA-GA-ELM

In the overall environment of economic globalisation and accelerating information networking,the level of logistics services has become an important criterion to measure the comprehensive economic capacity of a country and region.As the leader of the Yangtze River Delta city cluster,Shanghai is not only a domestic economic centre,but also an international economic and trade centre,and the development of its logistics industry has a strong radiation-driven effect on the Yangtze River Delta city cluster.It is of great theoretical and practical significance to forecast the logistics demand in Shanghai objectively and accurately.Based on the theory of regional logistics demand forecasting,this paper constructs a logistics demand forecasting index system for Shanghai based on the principles of regional logistics demand forecasting index selection and the actual development situation of Shanghai,and uses single factor fitting method and Spearman’s rank correlation coefficient to screen the relevant indexes for Shanghai logistics demand forecasting.The KPCA-GA-ELM forecasting model is constructed and compared with the Partial least squares model,BP neural network model,ELM model and GA-ELM model,and the superiority of the KPCA-GA-ELM model is finally verified.The grey system theory is introduced and a grey forecasting model GM(1,1)is established to forecast the impact indicators of Shanghai’s logistics demand.The forecast results are used as the input of the KPCA-GA-ELM model to obtain the forecast results of Shanghai’s regional logistics demand in the next six years.Through the research and analysis of logistics demand forecasting in Shanghai,the main research elements and conclusions are obtained as follows.(1)Based on the study of the literature related to regional logistics forecasting,the study analyses the actual development of Shanghai,combines the principles of selecting regional logistics demand forecasting indicators,selects freight turnover as a measure of Shanghai’s logistics demand and 15 logistics demand impact indicators,uses single-factor fitting analysis to clarify the law of action between the impact indicators,calculates Spearman’s rank correlation coefficient for the selected The 15 selected impact indicators were screened out,and those with low correlation were excluded,and a Shanghai logistics demand forecasting index system was finally constructed,comprising one measurement indicator and 14 impact indicators.(2)Based on the study of regional logistics demand forecasting methods,the study uses the KPCA algorithm to extract features from the impact indicators,and at the same time combines the GA algorithm with the ELM algorithm,uses the GA algorithm to optimise the initial random parameters of the ELM model,establishes the KPCA-GA-ELM model,and compares and analyses it with the Partial least squares model,BP neural network model,the ELM model and the GA-ELM model The KPCA-GA-ELM model has MRE=1.55%,RMSE=486.13,R2=0.9731,with stronger generalization ability as well as prediction performance,which affirms the value of the model.(3)This study analyses the change pattern of the impact indicators of logistics demand in Shanghai,introduces the grey system theory,selects the historical data of the impact indicators of logistics demand in Shanghai and establishes a grey forecasting model GM(1,1)based on time series,forecasts the data of each impact indicator of logistics demand in Shanghai from 2020 to 2026,and uses the forecast results as the input data of the KPCA-GA-ELM model to forecast the logistics demand in Shanghai in the next 6 years.The prediction results are used as input data for the KPCA-GA-ELM model to forecast the logistics demand in Shanghai in the next six years,which provides important data support and decision-making reference for the Shanghai government and enterprises in formulating scientific logistics development policies and planning the layout of regional logistics facilities.