- 作者: 段良雄
- 中文摘要: 一般皆認?較正確之函數指定可增加模式之可轉移性,本研究之主要目的即在於探討其影響程度。將臺南市之數據依地理界限分成南北兩區以進行都巿?模式之可轉移性分析。首先建立簡單線性函數型式之多項羅機(MNL)模式以做 為供比較之基本模式,再分別對各解釋變數做Box-Tukey轉換以建立較複雜函數型式之多項羅機與巢式多項羅機 (NMNL)模式。?究結果發現Box-Tukey轉換確可顯著增加模式之解釋能力與轉移能力,但巢式多項羅機模式之結果則不佳。
- 英文摘要: It is generally believed that a better specified model will have better transferability. It is the purpose of this paper to explore the degree of its effect. Using data from Tainan city and separating it into south and north areas according to its geographical boundary, this paper presents an intra-urban transferability analysis. Multinomial logit models with a simple linear functional form were built as base models for comparison. More complicated functional forms of explanatory variables were obtained through Box-Tukey transformation and multinomial logit and nested multinomial logit models were built with them. It was found that Box-Tukey transformation could significantly increase the explanatory ability as well as transferability of multinomial logit models. However, the results of nested multinomial logit models were not good.
- 中文關鍵字: Box-Tukey transformation; transferability; multinomial logit; nested multinomial logi; mode choice
- 英文關鍵字: --