臺灣博碩士論文加值系統

本研究主要在於探討如何結合顯示性偏好法與敘述性偏好法兩者於選
擇行為分析上的優點，而建立整合模式；並以台南-台北間之城際運輸運
具(包括火車、飛機、巴士及新運具高鐵等四種)偏好作為實證研究對象。
而從模式分析中本研究獲得以下幾點主要結論：1.顯示性偏好模式較能反
應旅運者之實際選擇行為，但卻 因研究方法之限制而缺乏新運具(高鐵)
之偏好資訊，因 此無法進行該運具之預測分析。2.敘述性偏好模式中之
慣性變數表現十分顯著，可知該模 式所代表的選擇行為甚受顯示性偏
好(即實際選擇行為) 的影響。3.顯示性偏好模式與敘述性偏好模式之間
存有尺度因子的 差異，而其所校估出之數值小於一，表後者之隨機誤差
項的變異程度大於前者。4.從整合模式之校估結果發現其值均是介於顯示
性偏好模 式與敘述性偏好模式兩者之間，即所建立之整合模式確 實能
對兩模式發揮截長補短的功效；而對於新運具轉移 運量預測之應用，整
合模式亦將是較顯示性偏好法與敘 述性偏好法為佳的選擇。

The purpose of this research is to propose a method which
capitalizes on the complementary strengths of RP and SP choice
data to obtain a joint model that can deal with new mode and
capture real market behavior accurately . This approach was
tested with intercity mode preference data collected in the
Tainan-Taipei corridor . This research explored some issues
related to RP , SP and joint models , and found the following
conclusions :1. The RP model had the best ability to reflect the
real market behavior , but could not predict the market
share of new mode.2. The inertia variables of SP model were very
significant , which implied that the SP choice model was
influenced by real choice behvior . In other word , the
state dependence was correlated between RP and SP models.3.
The scale factor''s problem did exist between RP and SP data ,
and the estimated value of the scale was less than one ,
which indicted that the variance of SP data was greater than
that of RP data.4. The joint model performed better than RP and
SP models in forecasting the new mode''s market share.