許添本臺灣大學：土木工程學研究所蔡佳佳Tsai, Chia-ChiaChia-ChiaTsai2007-11-252018-07-092007-11-252018-07-092005http://ntur.lib.ntu.edu.tw//handle/246246/50494摘要 我國歷年機動車輛數持續不斷增加，尤以機車增加情形為最多，此種以機車為私人運具發展導向的情形，亦在許多亞洲國家出現，而機車與汽車混合車流情形使得各種與機車相關之問題發生頻繁，機車持有數量之變動受到重視，因此影響持有之因素及未來機車數量之變化趨勢亦成為政府施政時之重要參考依據，本研究藉日本汽、機車數量之發展趨勢，對機動車輛發展進行階段性定義，以機車數量、汽車數量與國家機動性和經濟性發展指標進行探討，其中發現越南近年發展趨勢與台灣類似，應注意未來機車數量可能遽增之情形。 過去研究指出汽車與機車持有之間有其替代性存在，因此本研究以此為出發點，以聯立方程模式對國內台北市、台中市與高雄市三個地區進行探討，除探討汽車與機車持有替代關係之影響情形，同時進一步瞭解大眾運輸的發達與否，是否會對私人運具之持有造成影響。研究結果發現台北市汽車、機車與大眾運輸使用間彼此有相互替代性存在；台中市則是發現家戶汽車數量會影響家戶持有機車數量與使用大眾運輸人數，家戶機車數量亦會替代汽車之持有數量，家戶使用大眾運輸人數則間接透過家戶持有汽車駕照人數對汽車持有數量有所影響，並由影響因素發現，所得並非影響機動車輛持有之主要因素；高雄市則是汽機車間有直接替代關係，但持有機車或汽車均與所得無關。因此聯立方程模式可用以顯示機車持有、汽車持有及大眾運輸使用間之直接與間接影響關係，且其模式誤差較卜瓦松模式小。另以總體資料建立台灣地區汽車持有與機車持有模式，透過主成分分析方式解決變數間相關性過高而產生的共線性問題，並與一般迴歸進行比較，雖其誤差較聯立模式小，但兩種方式結果均屬優良。 關鍵字：機車持有、汽車持有、聯立方程式。Abstract The number of mobile vehicles in Taiwan, especially motorcycles, has been increasing for many years. The trend of using motorcycles for private communication has also been discovered in many other Asian countries. Since the mixed traffic flow causes more motorcycle-related problems, the fluctuation trend of motorcycle ownership has been looked into. Furthermore, factors that affect motor vehicle ownerships and future trend of total motorcycles are important bases for government policies. For this reason, this study defines different motorcycle ownership phases according to observed trends of motor vehicles in Japan. Each phase with numbers of motorcycles and cars represents a country’s mobility and economy indexes. The recent development trend in Vietnam is found to be similar to Taiwan and requires careful attention to the rapid increase of motorcycles. In past studies, car and motorcycle ownerships are found to be inter-replaceable. From this starting point, this study investigates on the relationship between cars and motorcycles in Taipei, Taichung and Kaohsiung by utilizing a simultaneous equation model. In addition, whether or not the development of public transportation will affect private transportation modes has been examined. The results show the replaceable relationship in a household between car ownership, motorcycle ownership and the number of public transportation users in Taipei. In Taichung, household car ownership would affect household motorcycle ownership and the number of public transportation users. The number of public transportation users is affected indirectly by the number of household car licenses. Furthermore, income is not the main factor for motor vehicle ownerships. On the other hand, direct relationship between car ownership and motorcycle ownership has been discovered in Kaohsiung. However, income does not affect household car and motorcycle ownership. The simultaneous equation model has shown the indirect and direct relations between motorcycle ownership, car ownership and the number of public transportation users and has proved to have fewer errors than a Poisson model. Lastly, solving the collinear problem between factors by principal component analysis has allowed data that are collected in Taiwan be aggregated into building a car and motorcycle ownership model. Although the MAPE of the regression is fewer in errors than the simultaneous equation model, both are within satisfactory ranges in the goodness test. Key Words：Motorcycle ownership; Car ownership; Simultaneous Equation Model目錄 第一章 緒論 1 1.1 研究背景 1 1.2 研究動機與目的 2 1.3 研究範圍 5 1.4 研究內容 5 1.5 研究流程 6 第二章 文獻回顧與比較 9 2.1 國內外汽機車持有課題文獻回顧 9 2.2 國內外文獻綜合整理 27 2.3 國內外汽機車持有模式方法整理 30 2.3.1 各類持有模式優缺點比較 34 2.3.2 國內外文獻方法應用模式之比較 37 2.4 影響汽、機車持有因素彙整 40 2.5 小結 42 第三章 國內外機車與小客車持有狀況分析 44 3.1 資料背景說明 45 3.2 機動車輛持有發展階段 46 3.3 亞洲各國發展趨勢 50 3.4 台灣地區三市之發展趨勢 54 3.5 小結 55 第四章 資料蒐集與變數處理 57 4.1 資料背景說明 57 4.2 資料彙整與處理 57 4.3 變數選擇 58 4.4 各變數基本統計量描述 60 4.5 聯立方程式基本介紹 62 4.5.1 聯立方程式模式之認定 63 4.5.2 聯立方程模式之參數校估方法 66 4.6 聯立方程模式假設 67 4.7 模式構建 71 第五章 汽機車持有模式比較分析 72 5.1 台北市聯立方程模式 72 5.2 台中市聯立方程模式 74 5.3 高雄市聯立方程模式 75 5.4 模式比較 77 5.5 台灣地區縱斷面聯立模式 79 5.6 建立聯立模式時應注意之要點 81 5.7 小結 82 第六章 結論與建議 85 6.1 結論 85 6.2 建議 86 參考文獻 87 附錄一 亞洲國家歷年機車、自用小客車持有率與MCrate趨勢圖 90 附錄二 問卷資料內容與選項編碼摘錄 97 附錄三 台北市、台中市、高雄市家戶橫斷面資料變數相關係數值 106 附錄四 台灣地區總體縱斷面資料變數相關係數值 107919384 bytesapplication/pdfen-US機車持有汽車持有聯立方程式Motorcycle ownershipCar ownershipSimultaneous Equation Modelthesishttp://ntur.lib.ntu.edu.tw/bitstream/246246/50494/1/ntu-94-R92521503-1.pdf