每個人在面臨特定的選擇集合時, 可能會有不同的選擇行為發生, 如果將每個人的反 應集合起來, 組成一組資料, 利用這些資料, 我們就可以導出一種行為模式。一般的 行為模式都將應變數視為連續變數( 例如需求量),近年發展出的個體選擇模式(Discr -ete Choice Model)則將應變數視為間斷的(Discrete)孌數( 例如交通工具的選擇: 汽車=1、火車=2、飛機=3等等) 。藉由這種模式, 我們可以進一步研究或預測人類在 面臨各種選擇時所產生的行為。 一個行為模式, 包含了可以被觀察的自變數和未知的參數, 我們利用抽樣觀察而得的 資料來推估這些未知的參數, 但不同的抽樣方法會導致不同的估計值。一般而言, 在 這類問題中應用到的抽樣方法有三種:1. 簡單隨機抽樣、2.外生分層抽樣、3.內生分 層抽樣。本文主要的研究內容集中在探討運用不同的內生分層抽樣法所推估的參數值 之間的差異。 研究方法大致可分為三個步驟: 一. 閱讀前人的文獻, 彙總整理前人的方法, 并發現新的抽樣與估計法: 此法是先決 定內生分層后每層的抽樣比例, 再用系統抽樣的方法抽出樣本, 最后用最大概似法對 參數加以推估。以Binary Logit Model為例: P(1︱x)=1/(1+exp(-xb)) P(0︱x)=1/(1+exp(xb)) 其log likelihood fuction為 logL=Σ {y [log(gl)+log(p(1︱x )]+(1-y )[log(g0)+log(P(0︱x )]-log[gl P (1︱x )+g0 P(0︱x )]} 其中y =1 or 0;gi是選擇i 選項的抽樣比例(i=1 or 0);x 為自變數;b是欲估計的參 數。 二、證明利用新法所產生之估計式, 和前人所研究出的估計式同樣具有一致性。 三、模擬一組母體, 再比較新法與前人的估計式間的差異。

近年來觀光產業的快速發展，使相關業者面臨激烈的競爭，旅遊目的地行銷的概念應運而生。隨著產業結構的轉型，地方都市無不透過觀光行銷來吸引投資者或遊客之進入，以增加自身競爭力，因應全球化下的競爭。 　　觀光意象是遊客對於一旅遊目的地所持的知覺、看法與印象的重組，其傳達遊客腦海中對旅遊目的地遊憩屬性的偏好，具有宣傳與行銷的功能。一般而言，觀光意象在遊客對目的地之選擇行為過程中扮演了重要的角色，具有良好觀光意象的旅遊目的地通常具有較高的吸引力，且正向的觀光意象對都市透過觀光收益來促進都市發展及經濟成長亦有顯著效果。觀光意象包含了遊客的認知與情感面，遊客將據其偏好選擇喜歡的目的地作為旅遊地點。然而，過去研究較少探討遊客主觀感受與其對地方之情感連結等心理層面對選擇行為之影響，據此，本研究以台北市貓空地區為例，以觀光意象作為前因變數，探討其對遊客之地方依附、旅遊品質、滿意度、選擇行為之影響路徑及各變數間之交互影響效果。 　　本研究以遊客問卷之設計，對非當地居民之遊客作為抽樣調查之對象，並以因素分析與結構方程模式作為資料分析方法，所得主要結果為：觀光意象對地方依附、旅遊品質與選擇行為有直接正向影響，而影響遊客對旅遊目的地選擇行為最重要的因素為地方依附。 / With the change of industry structure, tourism industry has become one of important industries in a city. Facing the competition under the global age, the concept of destination marketing has widely aroused attention because of fast development of tourism industry and keen competition among relative industry proprietors. Destination image is sum of beliefs, ideas and impressions that a person has of a destination. Destination image has been shown to be vital influence on travelers' travel behavior, that is, the urban with a positive and good image perceived by tourists can bring out the urban a significant effect on urban development and economic growth. A tourist's intent to visit a destination is determined by a combination of cognitive and affective image. Tourists usually develop emotional associations with destination where they're lovely to visit. However, very limited research has been conducted personal emotional or meanings tourists attach to the places they visited and experienced. Thus, the study takes Maokong Area of Taipei city for example, and the purpose of this paper is to examine the relationship among destination image, place attachment, trip quality, satisfaction, and traveling behavior. In research method, the survey was quantitative research oriented in order to understand visitors' ideas toward the destination images of Maokong area. Both factor analysis and structural equation modeling were used to analyze for the study. The major results of this research were found that destination image directly has positive influence on place attachment, trip quality, and traveling behavior, and place attachment is the most important cause to influence tourists' behavior.

目前區段徵收乃是政府進行土地開發的方式之一，多用來取得所需之公共設施用地或特定目的事業用地，高鐵桃園車站特定區即採區段徵收方式進行開發，以取得高鐵所需之車站用地。由於現行區段徵收為抵價地式制度，被徵收之地主有領取抵價地與領取現金補償兩種選擇，因此，政府實施區段徵收時，將受到地主不同選擇行為之影響而有不同的辦理成效，使得區段徵收地主選擇行為在徵收辦理過程中，具有決定性的影響力。 　　本研究自土地開發之觀點探討區段徵收地主選擇行為，並著重於選擇行為影響因素之分析。由於相關文獻針對此部分的研究甚少，因此，首先歸納區段徵收地主選擇行為之影響因素，共分為地主屬性因素、政策管制因素及未來期望因素三類，並以高鐵桃園特定區為實證案例，透過問卷調查與Logit模型之建立，分析各項因素對特定區地主選擇行為的影響程度；其次則分析選擇行為的影響效果，以了解區段徵收地主選擇行為之前因後果，並進一步提出政策建議。 　　經由高鐵桃園車站特定區之實證分析，本研究得出以下六點結論： 　　1.大地主傾向領取抵價地，小地主與共有地主傾向領取現金補償，可藉由土地屬性之掌握，預知地主可能的選擇行為。 　　2.公告現值調漲時機關係徵收補償費之支出，並左右地價與市場景氣對選擇行為的影響，政府應先公告實施區段徵收後才發布都市計畫，以減少徵收補償費，提高財務可行性。 　　3.區段徵收地主選擇行為受價值觀之影響大，應考量地主保有祖產價值觀等個人屬性之影響，以預知選擇行為。 　　4.建築基地最小開發規模攸關小地主配地權益與土地開發型態，政府應考量土地規模與使用，訂定合理的建築基地最小開發規模。 　　5.抵價地領回比例與分配區位受公共設施用地規劃之影響，為領地地主改領現金補償的主要原因，可藉由抵價地相關規定誘導地主選擇行為之改變。 　　6.政府辦理區段徵收時，普遍未充分告知計畫內容與相關規定，影響地主權益。 / Zone expropriation is one kind of land development methods through which government can obtain the land for public facilities or specific uses. For example, the development of Taoyuan High Speed Rail station special zone is of this kind. According to zone expropriation system, there are two compensatory choices for landowners. One is by cash and the other is by pay-for-land. Therefore, landowner's choice behavior would play an important role in the process of zone expropriation. 　　With the view of land development, this paper discusses landowner's choice behavior in the process of zone expropriation and focuses on the factors of choice behavior. Here we sum up three kinds of factors that will affect landowner's choice behavior: they are landowner attributes, government regulations, and landowner expectations. According to the empirical study of zone expropriation in Taoyuan High Speed Rail station special zone, this paper has six conclusions: 　　1.Large landowners prefer to choose pay-for-land. On the contrary, small and common landowners prefer cash compensation. 　　2.To reduce the amount of cash compensation, the government should announce zone expropriation more early than urban plan. 　　3.The concept of keeping ancestral estate would influence landowner's choice behavior very much in the process of zone expropriation. 　　4.The development scale would affect the right of small landowners and the form of land development. The government should set a suitable development scale by considering the scale and use of land. 　　5.The assigned proportion and location of pay-for-land is the main reason why landowners with pay-for-land change to choose cash compensation. 　　6.The government seldom informed landowners of the whole plan and the relevant regulations in the process of zone expropriation.

政府辦理區段徵收希望帶動新社區的繁榮，達成都市整體發展，取得公共設施用地及節省龐大建設經費支出等之目標，而地主希望藉由政府的開發行為，享有土地利用價值提高、公共設施完善等好處，然而所謂的公私互蒙其利，共創政府與民眾雙贏的開發方式，卻仍常產生地主為了財產權保障抗爭之情形而導致區段徵收作業延宕，因此，本研究以高鐵新竹車站特定區區段徵收案為例，針對地主選擇行為進行探討，透過群落分析歸類地主選配抵價地的型態，經由因子分析找出影響選配抵價地行為之因子，最後藉由羅吉特建立地主選配抵價地模型，希望可以提供政府部門及地主在選配抵價地之參考。研究結果如下： 一、 透過群落分析，予以歸類分析，劃分易於解釋的分配類型，可獲得二個群組四個方案，由此可瞭解地主選配抵價地型態。 二、 參與抵價地配地作業以男性地主居多，以年齡層多集中於中年左右，受訪地主多為已婚；受訪地主針對各項選配抵價地之影響因素均表示影響程度上都是以影響很大為居多，其中受訪地主多數認為使用分區類別、面前街道寬度、鄰近嫌惡設施程度、坵塊形狀、街廓評議地價高低等因素對於選配抵價地影響很大，又以鄰近嫌惡設施程度為居高。 三、 經由研究得知目前政府提供予地主之資訊管道不足或宣傳不周，對於整體區段徵收作業仍需加強與地主資訊的傳遞，以及充分公開整體計畫內容。 四、 部份地主不太滿意先抽籤再選配街廓的抵價地配地作業方式，其原因有主要為因籤號先後導致無法選配期望街廓佔多數，其次為權利價值較小者可能無法選配街廓，而需再與他人合併分配。 五、 建構八種類型之二項羅吉特模式，並透過AIC及SC值找出能力較佳的群組方案，由方案中得知地主多數還是以居住適宜性及小規模開發為主要。 六、 由二項羅吉特模式實證結果顯示，模型共通變數為土地變數的使用分區類別、基地最小開發規模訂定、鄰近嫌惡設施程度、緊臨遊憩設施狀況、坵塊座向，以及地主行為變數面前街道寬度影響程度、鄰近嫌惡設施影響程度、區段徵收前土地座落位置影響程度、性別、學歷、職業等等，均顯著影響地主選配抵價地之行為。 / Zone-expropriation by government hopes to achieve these targets like exciting new community's prosperity, achieving the overall development of the new city, accessing the public facilities and saving huge expenditures. By the government's development activities, the landlords enjoy the benefits of increased value of land and good public facilities. Handling Zone expropriation creates a win-win situation between the Government and the landlords. However, in the process of Zone-expropriation often generate opposition of landowners to protect their property rights which led to delays in Zone-expropriation operations. Therefore, this study of THSR (Taiwan High Speed Rail) Hsinchu Station Zone- expropriation as an example and discuss the landowner’s choice behavior. Through cluster analysis classified the landowner’s choice behavior. And by factor analysis to identify the major factors of landlord’s choice behavior. Finally by Logit modeling the landlord model hopes to provide some valuable suggestions for government. The study results are as follows: 1. Through cluster analysis, the study classified and divided into two groups that are easy to explain. The two groups contain four plans each and they can explain the landlord’s choice behavior. 2. Landlords of participation of Zone-expropriation are mostly male with age around 40 to 60 years old. These mostly married landlords think these are the major factors, the use of zoning category, the street width, the neighboring facilities, the shape of street block, the value of street blocks. The neighboring facilities rank the top 1 factor. 3. Through the research shows that government lacks of information or promotional channels. Government should improve the communication with landlords and fully disclosed the contents of the overall Zone-expropriation plan. 4. Landowners are not satisfied by the flow of ballot at first and then land choosing. The major reason is landlords can not choose the expected lands after ballot. And the second reason is the land loads with smaller land value can not choose street blocks unless they combine with other landlords’ land value and choose the expected land. 5. In the study, construction of eight types of models from two of Logit modeling, through the AIC and SC values come out the better group and plan. It shows the better plan is the small-scale residential development and good environment for the most landlords. 6. By the empirical results from Logit model show that model common variables are the land use zoning category, the base size of the minimum development, near the level of aversion facilities, recreational facilities close to the zone, street block location, the width of the street, the impact of land location before zone expropriation, gender, education, occupation, etc. They are significantly affected the selection behavior of land choice.