以二元羅吉特模式探討大型航空樞紐城市於空間與時間可移轉性之研究

Application of Binary Logit Model on Temporal and Spatial Transferability of Large-Hub Primary City

郭仲偉(Chung-Wei Kuo)

《中國土木水利工程學刊》 20卷2期 (2008/06) Pp. 241-256

https://doi.org/10.6652/JoCICHE.200806_20(2).0008

摘要

航空樞紐的重要性除了能帶動所在城市的經濟持續發展之外，同時也是整個國家發展的重要基礎。而一個大型航空樞紐城市的形成，必有一股強大的力量在推動，除了主觀意識影響之外，也有該城市先天客觀條件的優勢造就其發展。因此分析何種因素能夠解釋大型航空樞紐城市形成，為本研究所要探討課題之一。本研究以美國為研究基礎對象，利用二元羅吉特模式分析其大型航空樞紐城市的客觀條件影響因素。而結果顯示，除了都會區人口條件以及經濟所得條件是影響大型樞紐城市形成原因之外，觀光資源也是主因之一。最後透過模式更新方式進行時間與空間上之實例驗證，並利用可移轉性統計檢定方法加以判斷，模式是否具可移轉性。結果顯示，在時間移轉部份，以結合移轉性估計法或貝氏更新法移轉效果較佳，在空間上移轉，則以修正方案特定常數與參數尺度法最佳。本研究結果可提供相關單位在研究未來大型航空樞紐城市時之參考。

關鍵字

大型航空樞紐城市 ； 二元羅吉特模式 ； 結合移轉性估計法 ； 修正方案特定常數與參數尺度法 ； 貝氏更新法

並列摘要

Air hub can not only stimulate the consistently economical development of the city where the hub is located, but also is the foundation of whole national development. A formation of a large-hub primary city certainly has an enormous power pushing behind. In addition to the influence of subjective consciousness of the city itself, it also has some advantages to speed this formationt. It is therefore our main purpose to investigate factors influencing the formation of large-hub primary city in the United States by using binary logit models. The resulst show that metropolitan population and Gross Domestic Product are two significant factors that influence the large-hub primary city formation. Tourist resource also is one of the main factors. Finally, updating techniques are carried out to perform the investigation of time and space effects, and measures of transferability are used to judge whether models can be transferred either in time or space. The results show that for temporal transferability, the best method is combined transfer estimation and Bayesian updating. For spatial transferability, the best method is corrected alternative specific constants and the parameter scaling. The results obtained from this study can provide a reference on how largehub primary city is forming.

並列關鍵字

large-hub primary city ； binary logit model ； combined transfer estimation ； updating alternative specific constants and scale parameter ； bayesian updating