Application of Regression Analysis and Time Series Models for Freight Demand Estimation and Forecast at Lianyungang Port
Shuang Tang,Sudong Xu,Jianwen Gao,Mengdi Ma
DOI: https://doi.org/10.1061/9780784482292.246
2019-01-01
Abstract:Estimating and forecasting the inbound and outbound freight demand over time is important for port planning. This paper focuses on the analysis of port cargo structure and the forecast of the inbound and outbound freight demand of Lianyungang Port. By analysis, coal and metal ore are key types of goods affecting inbound freight, and as for outbound freight is metal ore. ARIMA model, Holt-Winters model, and multiple linear regression model are applied to evaluate the freight demand based on the datasets from 1990 to 2016. Social-economic indicators are taken into account, including gross domestic product, total fixed assets investment, disposable income of urban residents, total industrial output value, total import foreign trade, total export foreign trade, and the city's financial revenue. The results show that the ARIMA model and Holt-Winters model are selected to be optimal models to predict the inbound and outbound freight volume for 2017-2019, respectively.