查詢結果分析
來源資料
頁籤選單縮合
題 名 | 整合顯示偏好與敘述偏好數據的運具選擇模式=Mode Choice Models Combining Revealed Preference and Stated Preference Data |
---|---|
作 者 | 段良雄; 王郁珍; | 書刊名 | 運輸計劃 |
卷 期 | 28:1 1999.03[民88.03] |
頁 次 | 頁25-59 |
分類號 | 557.16 |
關鍵詞 | 新運具運量; 敘述偏好; 顯示偏好; 整合模式; 多項羅機; 巢式羅機; New mode ridership; Stated preference; Revealed preference; Combined model; Multinomial logit; Nested logit; |
語 文 | 中文(Chinese) |
中文摘要 | 本文之主要目的在於討論新運具的運量預測問題。利用自行設計的問卷所蒐集的 顯示偏好與敘述偏好數據共建立了三類運具選擇模式:1.顯示偏好模式,2.敘述偏好模式,3. 結合兩種數據的整合模式。臺南臺北間的域際大眾運輸運具選擇問題為實證研究對象,顯示 偏好模式包括火車、飛機、與巴士三種運具,敘述偏好與整合模式則另加入新運具高速鐵路。 本研究共建立了三種旅次的模式,包括了非公商務旅次模式,公商務旅次模式,與將兩者合 併的全部目的旅次模式。所建立的運具選擇模式有多項羅機模式與巢式多項羅機模式。實證 研究的結果顯示下列結論:1.顯示偏好模式之解釋能力不錯,但因缺乏新運具之選擇資訊, 故難以進行新運具之運量預測。2.敘述偏好模式之狀態相依相當顯著,亦即此模式受前期的 實際選擇行為相當大的影響。3.顯示偏好模式與敘述偏好模式之間存有尺度因子的差異,而 後者之隨機誤差項的變異程度大於前者。4.整合模式之解釋能力良好。本文亦探討了敘述偏 好模式的情境組合的順序對偏好的影響、整合模式的校估方法、以及利用整合模式預測新運 具運量的方法。 |
英文摘要 | This paper discussed the prediction problem of new mode. Using revealed preference and stated preference data collected through self-designed survey, three types of models were built. There were revealed preference models, stated preference models, and combined models using both data. The intercity public transportation mode choice problem was empirically studied. The modes used in the revealed preference models were rail, air, and bus. The new mode, high speed rail, was added to the stated preference and combined models. Three different trip models were studied, i.e., non-business trip models, business trip models, and all-purpose models. The choice structures used were multinomial logit and nested multinomial logit. The empirical results showed the following conclusions. 1. The revealed preference models had relatively good explanatory ability but were not able to predict the market share of new mode due to data shortage. 2. The state dependence phenomena were significant in the stated preference models. This meant that these models were affected by previous choice behavior. 3. There was a scale difference between revealed preference and stated preference data. The latter had a larger random error variance. 4. The explanatory power of combined models were relatively good. This paper also discussed the effect of transportation scenario's sequence on stated preference models, the calibration method of combined models, and the method to predict new mode's market share. |
本系統中英文摘要資訊取自各篇刊載內容。