本研究目的在於探討兩岸航線選擇行為之方案相關性。以往處理方案相似性問題大多使用巢式羅吉特模式,但此模式僅能處理一部分的相似問題,為深入探討方案之相似性,本研究首先使用一般化巢式羅吉特模式,此模式可探討多個替選方案間之相關性,以及推測部分重覆的組合,藉此探討不同巢層內方案之相似性;接著運用誤差成份羅吉特模式之特質,找出方案內之誤差成份項相似且重疊之特性;藉以分別探討各模式替選方案間之方案相似性問題,最後使用預測指標、彈性分析與政策模擬等,進行模式間全面性之比較與驗證。實證結果顯示,於兩岸直航航線之選擇模式下,最佳模式為加入隨機參數的誤差成份羅吉特模式;在政策模擬測試結果之下,同時減少接駁時間與飛機票價,可顯著地提升航線市佔率。
The purpose of this study is to analyze the choice behavior for cross-strait flights between Taiwan and China, and the alternative similarity across four flying routes. The most common model used in studying alternatives is the Nested Logit model. Such model is used to conduct with the problem of alternative similarity. But it can only help to deal with part of the alternative similarity. In order to go deep into the issue, this study adopts two advanced approaches to solve it. First, Generalized Nested Logit model is chosen to explore multiple correlations between alternatives and to speculate repeated alternatives across different nest structures. Then, Error Components Logit model is chosen to find out error components between different alternatives and their overlapping features in order to explore the alternatives similarities. In addition, prediction index, elasticity analysis and policy simulation methods are used to compare and verify the foregoing choice models. The empirical results show that the Error Components Logit with random parameter has the best explanatory power and acceptable predictive ability. The results of policy simulation show that improving the level of access time to departure airport and adopting price promotion can raise significantly the market share for some specific flying routes.