貨櫃吞吐量預測模式之比較研究

A Comparative Study of Forecasting Models on the Prediction of Container Throughput Volumes

本研究之目的為使用六種單一變數預測方法預測台灣國際港口之每月貨櫃吞吐量，六種預測方法包含有古典分解法、三角函數迴歸、季節性虛擬變數、灰預測、組合灰預測及SARIMA。研究對象為台灣地區三大國際港埠(基隆港、台中港與高雄港)之貨櫃吞吐量，研究中利用實證分析，以驗證何者可提供預測最佳之精確度，經由利用平均絕對誤差(MAE)、平均絕對誤差百分比(MAPE)及殘差均方根(RMSE) 等評估指標比較後，發現基隆港以古典分解法，而台中港與高雄港以SARIMA 的預測能力最佳。本研究之結果可提供港埠當局未來進行港埠規劃時之參考。

The purpose of this study is to provide a more accurate prediction model on the container throughput for rendering a reference to authorities. Six different univariate methods, namely the Classical Decomposition Model, the Trigonometric Model, the Regression Model with Seasonal Dummy Variables, the Grey Forecast, the Hybrid Grey model, and the SARIMA, have been used. The contribution of this research is to compare the forecasting results of the six univariate methods based on commonly used evaluation criteria, MAE, MAPE and RMSE. We found that, the Classical Decomposition model is a reliable prediction method for forecasting Keelung port container throughput, and the SARIMA is the best method for Taichung and Kaohsung port. The outcome of this work can be helpful to predict the near future demands for the container throughput of the international port.