关键词: 天津港, 港口吞吐量, ARIMA, ARIMAX-SVR

Abstract: Accurate prediction of port cargo throughput is of great significance to the development planning of ports. In order to improve the prediction accuracy of port cargo throughput, a combined prediction model based on ARIMAX-SVR is established. Taking Tianjin Port as an example, the data of cargo throughput from 1999 to 2018 are selected as the research object. BP neural network is first used to supplement the missing data, and then the main factors affecting cargo throughput are selected through the Pearson correlation analysis. Then, based on the ARIMA model, a multivariate time series prediction model is established. In order to further improve the accuracy of the model, the multivariate time series prediction model modified by the support vector regression model is finally introduced. Through empirical analysis, the ARIMAX-SVR combination model can be obtained with higher accuracy and better prediction effect than the ARIMA model, which is more suitable for port throughput prediction. The combination model established in the case of a few similar studies has some advantages.

Key words: Tianjin port, port throughput, ARIMA, ARIMAX-SVR