Research On Air Cargo Demand Forecast Based On Combination Model

Air freight,as an important component of modern logistics,combines the characteristics of large span and high complexity of the logistics industry,as well as the high investment,high technology,and high risk of the civil aviation industry.Under the influence of factors such as economic globalization and the rapid development of ecommerce,China’s international and domestic trade volume has significantly increased.The active trade exchanges,accompanied by industrial upgrading and consumption upgrading,have led to rapid development of the air cargo industry.Scientific forecasting of air cargo demand is crucial for formulating infrastructure planning and economic development strategies.This thesis takes the demand for air cargo as the research object.Firstly,literature research and inductive analysis methods are used to summarize the existing research on air cargo demand prediction indicators.Secondly,by analyzing the influencing factors,indicator construction principles,and methods of air cargo demand,relevant indicators that affect air cargo demand are preliminarily selected,and the low correlation indicators are removed using grey correlation analysis method to establish the final indicator system.Subsequently,based on the basic theoretical review of commonly used logistics demand prediction methods,principal component regression models,multivariate grey prediction models,and BP neural network models were selected from three different prediction types:Causal analysis prediction,grey prediction,and machine learning prediction to predict air cargo demand.Based on the basic principles of combination prediction,the prediction data of three single prediction models were used as input,and these three models were weighted and combined using the Shapley value method to construct a combination prediction model.At the same time,the effectiveness of the combination model was verified in empirical research.Then,combined with the GM(1,1)model,the air cargo throughput of Chongqing in the next five years was predicted,and relevant suggestions were proposed based on the predicted results.Finally,through empirical summary,this thesis found that the resolution coefficient of grey correlation analysis cannot be determined based on convention and experience when selecting air cargo demand prediction indicators.Based on the actual situation,this thesis further analyzes and verifies the rationality of the indicators by reducing values to improve resolution.However,considering the complexity of air freight,alternative solutions have also been proposed in the outlook.Overall,the research on the indicator analysis and model construction of air cargo demand prediction in this thesis can provide certain reference value for the government,airlines,and other scholars,and also provide some suggestions for the construction of air cargo hubs in Chongqing.