應用灰色理論於港埠貨櫃吞吐量預測全球貿易之研究

AN APPLICATION OF GREY THEORY TO GLOBAL TRADE PREDICTION BY PORT CONTAINER THROUGHPUTS

國際貿易為全球經濟景氣之表徵，本研究利用與全球貿易極為相關之港埠貨櫃吞吐量，進行灰色理論預測模式之建構。首先由灰關聯分析中得知，與全球貿易出口值之灰關聯度最高的前4大港，分別為布萊梅港、寧波港、新加坡港與天津港。然後利用該4港之貨櫃吞吐量與全球貿易出口值，進行GM（1,5）預測模式之建構。經由平均誤差百分比絕對值之評估，預測能力可達97.97%，Theil’s U統計量值幾近於0，顯示模式之預測效果頗佳，經與GM（1,1）模式比較亦良好許多，此一模式應頗具參考價值。

International trade growth is one major sign in the development of global economies. By analyzing port container traffic, this study attempts to construct prediction models of yearly global export trade value with grey theory. First, candidates among leading ports were selected using grey relational analysis: Bremen, Ningbo, Singapore and Tianjin ports were picked as each had higher grey correlations with global export trade value. Second, a GM (1. 5) model was projected in order to forecast trade values by analyzing container throughputs of these four ports. Results reveal that this model has well-forecast abilities with 97.97% of mean absolute percent error and almost zero of Theil’s U statistics. It was also better than that of GM (1.1).