本研究主要是希望針對採市場資訊的Merton Model、利用財務比率為變數的Logistic Model及混合用市場資訊及財務比率資訊的Hybrid Model做區別力的比較；且觀察出考慮提前違約是否能增加區別效力。進而將模型細分為: （一）以Merton Model為基礎所估出的風險中立下違約機率及客觀違約機率，及尚有為了修正Merton Model公司只會在負債到期日時才發生違約的基本假設的首次通過模型。 （二）採用Altman Z-Score所採用的五個判定公司的財務比率作為Logistic Model的基本變數。 （三）模型五、六、七除了利用Altman Z-Score所採用的五個判定公司的財務比率作為Logistic Model的基本變數。尚將風險中立下違約機率、客觀違約機率及首次通過模型違約機率分別當作的另一個解釋因子，形成同時結合市場資訊及財務資訊的Hybrid Model。 同時，採用了判別隨機性的Chi-Square Test、解釋力的Pesudo R2、區別力的S-K Test及ROC曲線、正確性的Brier Scores，此五個指標作為做為樣本內及樣本外模型的驗證方法。 經由驗證，得到的結論是同時包含市場資訊及財務資訊的模型較佳，且有考慮公司於負債到期日前就發生違約的因素對預警模型的判別效力有提高的正向作用。

處理多重事件或再發事件之事件發生時間的資料時，常會以Cox模式為基礎而予以延伸，其中較適合再發事件的模式為：A-G模式、GT-UR 模式、PWP-CP 模式及PWP-GT模式。這些模式又可按照是否以發生次數為分層變數，而分為未分層模式(包含A-G模式、 GT-UR 模式)，及分層模式(包含PWP-CP模式、 PWP-GT 模式)。 本論文將以改良的Cox延伸模式，包括對變異數進行修正或加入事故傾向(或隨機效果)，探討公務人員升等的快慢與哪些變數有關。變異數修正方式利用穩健標準誤以解決事件之再發時間之間的相依問題；事故傾向模式則主要是以隨機效果代表無法觀察到的個體間之異質性，且同一個體的各次發生時間共享相同的異質性，並假定異質性服從某種特定分配。對於各種Cox的延伸模式，我們可比較採用穩健變異數與否對估計及推論結果的差異，以及事故傾向加入前後，估計及推論結果與模式配適上的差異。 由本論文對公務人員升等資料的分析可發現，採用變異數修正方法時，未分層的模式有較小的變異數估計值，所以顯著的變數較多，包括性別、官等、教育程度及年齡；分層模式中顯著變數則只有官等及教育程度。若假定事故傾向服從對數Gamma分配，並加入於上述四種模式中，則顯著的變數與未加入事故傾向時一致，且各模式之下均無法拒絕所有人的事故傾向同為0的假設。這種現象或許是因為我們無法取得教育程度與公務人員考試及格種類之歷史資料，也有可能是因為公務人員升等的體制健全，且法規制定嚴謹，運作也有正常的模式可循所致。

自1978年中國開始經濟的改革開放政策以來，吸引外資便成為其推動經濟成長的重要手段之一。1992年鄧小平南巡，再一次宣示建立「社會主義的市場經濟體制」後，由於開放的經濟政策更為明朗，使越來越多的國家到中國直接投資。自此，中國成為全世界最受歡迎的外資投資國之一。甚至在2002年超越美國，成為全世界最大的外資接受國。如此多的國家對中國展開大規模的投資，其投資的規模與投資地區的選擇卻有相當大的差異。因此，本文的研究目的，將以1993至2003年世界各國對中國投資的追蹤資料（panel data），搭配固定效果模型（fixed-effect model）與隨機效果模型（random-effect model）的估計，並以目前中國沿海發展具代表性的北京、上海與廣東三個省（市）之比較，探究影響世界各國對中國直接投資的決定因素。 實證結果發現影響外商直接投資北京、上海與廣東的決定因素中，顯著影響的變數為相對工資率、對中國貿易依存度、相對匯率以及相對借貸成本。而其餘的變數，如相對國內生產毛額、相對每人國內生產毛額和相對國家風險等皆不顯著，反映了其皆非外商直接投資中國時所考量的決定因素。 / Since 1978, China has adopted the so-called “open door policy”, attracting foreign direct investment (FDI) has become one of the most important methods to facilitate its economic growth. However, foreign countries didn’t invest large amount toward China until Deng Xiaoping’s southern trip in 1992. The purpose of this study is thus to investigate the determinants of FDI from different foreign countries to different regions in China after foreign countries started to invest tremendous amount toward China. This study adopts fixed-effect model and random-effect model to investigate the determinants of FDI in China with panel data of Beijing, Shanghai and Guangdong during the period of 1993-2003. The result shows relative wage rate, trade dependence to China, relative exchange rate and relative borrowing cost are the most important factors in attracting FDI in Beijing, Shanghai and Guangdong during 1993-2003. Finally, in order to reduce the mistakes occurred in positive models and enable the study more rigorous, then uses more methods to test the models and the result.

本文採用 Chib， Nardari， 與 Shephard(2006) 的多變量動態因子隨機波動模型(MSV)， 來探討美、日、台三國的資訊、電腦類股股價報酬率波動的共同行為。 我們將模型中的因子解釋為產業的前景或信心，並藉由模擬的方式描繪出其樣貌，進而希望了解產業景氣循環在股價的波動行為中扮演什麼角色。 研究財務市場間的關聯性一值是一項重要的課題，也發展出各種的模型來描述既有的現象。 MSV 模型將看不到的解釋變量數量化，並將變數的波動行為切割為可由因子所解釋與不能解釋的部分。 且藉由將觀察值的誤差項以及單一因子的波動行為設定為隨機波動，放寬共變數變異數矩陣為定值的假設，讓每一時點都能依時變動，在同類的模型中對資料的設定是較少的。 在實證分析中我們有幾點發現：1. 因子能夠解釋資產間的波動行為，其反映在扣除因子波動之後的自有波動，其波動水準值的降低。 2. 在股價波動劇烈期間，因子解釋能力提高。 3. 因子的解釋能力在不同的國家中差異幅度很大，日本有超過一半的波動可以為因子的波動所解釋，而因子在台灣股價的波動行為只有兩成左右的解釋能力。

本文建構一個小型開放的經濟體系，價格具有僵固性的 DSGE 模型來探討提早釋出的關稅調降訊息所造成總體經濟面的影響。在政府簽訂貿易協定來降低關稅的訊息藉由媒體釋出時，同時民眾具理性預期的假設下，家計單位會因這樣的訊息而改變對未來的預期，在關稅還未實際調降時便改變行為決策。我們研究在不同的協商可能結果: (1)談判破裂，貿易協定未如預期簽訂。(2)談判成功，但貿易協定簽訂的結果比當初訊息所透露的降得更多或較少。(3)談判成功，並完全實現當初訊息的內容。(4)簽訂的時間高於預期，導致實行的時間延期。從結果中我們發現，提早釋出的關稅訊息會抑制民眾的消費和投資，在短期會立即造成需求面的負向衝擊。而降低關稅所帶來的市場活絡，會等到真正調降的時後才出現。所以越早釋出關稅調降的訊息，會造成經濟體系所需付出“等待政策實行的成本”越大。 / There is global trend of economic integration across the world by removing the trade barrier. While the free trade agreements normally include the tariff reduction, the negotiations of the agreement may take a long time, and in some cases, the negotiation may fail. Therefore, the tariff reduction’s effects on the economy can be different if it is realized as expectation or not. With a small open economy DSGE model, this paper examines the effects of news preannouncement on tariffs. With the assumption of rational expectation, households will change their expectation when the news of tariff decrement is preannounced. However, whether or not the news on tariffs can be realized as expectation will lead to different dynamics. In this study, we consider various plausible scenarios: (1) If negotiation fails, thus the news on tariff reduction is not realized (2) If negotiation succeeds, but the amount of actual decrement is more or less than people originally expected. (3) Negotiation succeeds and the content of news is fully realized. (4) Negotiation succeeds, but it takes more time than expected to be realized, thus the policy implementation is postponed. Our study reveals that the news preannouncement restrains consumption and investment before the negotiation is completed, and leads to a negative impact on the economy in the short run. The benefits of the tariff decrement appear if the policy is actually implemented as expected. However, the economy suffers more if the negotiation takes too much time after the news is released.

現今許多網路商店或電子商務將產品銷售給消費者的過程中，皆使用推薦系統的幫助來提高銷售量。如亞馬遜公司(Amazon)、Netflix，深入了解顧客的使用習慣，建構專屬的推薦系統並進行個性化的推薦商品給每一位顧客。 推薦系統應用的技術分為協同過濾和內容過濾兩大類，本研究旨在探討協同過濾推薦系統中潛在因子模型方法，利用矩陣分解法找出評分矩陣。在Koren等人(2009)中，將矩陣分解法的演算法大致分為兩種，隨機梯度下降法(Stochastic gradient descent)與交替最小平方法(Alternating least squares)。本研究主要研究目的有三項，一為比較交替最小平方法與隨機梯度下降法的預測能力，二為兩種矩陣分解演算法在加入偏誤項後的表現，三為先完成交替最小平方法與隨機梯度下降法，以其預測值對原始資料之遺失值進行資料插補，再利用奇異值分解法對完整資料做矩陣分解，觀察其前後方法的差異。 研究結果顯示，隨機梯度下降法所需的運算時間比交替最小平方法所需的運算時間少。另外，完成兩種矩陣分解演算法後，將預測值插補遺失值，進行奇異值分解的結果也顯示預測能力有提升。 / Recommender system has been largely used by Internet companies such Amazon and Netflix to make recommendations for Internet users. Techniques for recommender systems can be divided into content filtering approach and collaborative filtering approach. Matrix factorization is a popular method for collaborative filtering approach. It minimizes the object function through stochastic gradient descent and alternating least squares. This thesis has three goals. First, we compare the alternating least squares method and stochastic gradient descent method. Secondly, we compare the performance of matrix factorization method with and without the bias term. Thirdly, we combine singular value decomposition and matrix factorization. As expected, we found the stochastic gradient descent takes less time than the alternating least squares method, and the the matrix factorization method with bias term gives more accurate prediction. We also found that combining singular value decomposition with matrix factorization can improve the predictive accuracy.

本文採用一階段隨機邊界分析方法，建立一多投入多產出之投入距離函數模型，衡量民國81~91年間台灣66家國際觀光旅館經營之相對技術效率，同時探討造成不效率之因素。並利用投入導向Malmquist生產力指數進一步分析國際觀光旅館產業生產力改變的原因與幅度。 整體而言，台灣國際觀光旅館產業之經營效率大致呈現緩慢進步之趨勢。造成整體產業生產力提升之原因，主要為業者對於投入產出配置使用之效率改善，而非生產技術進步。除了整體經濟景氣因素當然影響觀光產業之興衰外，個別國際觀光旅館業者之效率表現仍受到許多因素影響。諸如規模、國際化程度等內部原因，及旅館是否位處風景區、所在地縣市政府對觀光發展投注之資源預算和該地市場競爭程度等外在因素。 從政府政策方面來分析，若以整體國際觀光旅館產業經營效率來衡量我國觀光產業興衰，民國87年隔週休二日政策及民國90年實施之全面週休二日政策，的確有效帶動國人休閒旅遊之風潮，促進國內觀光發展。而各縣市政府觀光相關支出與國際觀光旅館生產力變動之關係，呈現正向相關，則表示政府支出增加有助於觀光產業蓬勃發展，並增加國際觀光旅館之經營效率。由於觀光產業生產具外部性，政府積極利用觀光產業以帶動經濟景氣繁榮成長的政策應當可行。 / In this paper a multi-output, multi-input and input oriented distance function is built by one-stage stochastic frontier approach (Battese and Coelli, 1995) to estimate the relative efficiency of the 66 international tourist hotels in Taiwan in 1992~2002 and to explore the determinants of technical efficiency. In addition, the Malmquist productivity index model is used to analyze the range and the causes of the productivity changed. The results reveal that managerial efficiency of international tourist hotel industry improves gradually. The productivity growth can be attributed to efficiency gains, but not the effects of technical progress. There are significant differences in efficiency due to difference not only in the macroeconomic situation, but also in many other individual factors, such as scale, joining an international chain, located in scenic area, local government other economic service expenditure and competitiveness. Analyzing the government policy, the implementations of two-day-off twice a month policy in 1998 and two-day-off per week policy in 2001 have fostered domestic traveling and expanded tourist hotel industry. The local government other economic service expenditure has positive effect on efficiency of international tourist hotels. That is to say, increasing in local government tourist expenditure will promote tourism market and improve efficiency of a hotel. Because of the production externality in tourism industry the idea that government tries to promote tourism to benefit economic development would be workable.

本文於確定提撥退休金制度下，探討基金經理人如何決定最適資產策略規避薪資所得及通貨膨脹之不確定風險，求得期末財富效用期望值極大化。本研究首先擴展Battocchio與Menoncin (2004)所建構之資產模型，我們不僅探討來自市場之風險，同時考量薪資所得、通貨膨脹與費用率之不確定性，研究其對最適資產配置行為的影響，建構隨機控制模型，以動態規劃方法求解Hamiltonian方程式，研究結果顯示，我們可利用五項共同基金分離定理來描述投資人之最適投資決策：短期市場基金、狀態變數避險基金、薪資所得避險基金、通貨膨脹避險基金與現金部位。數值結果顯示，股票持有部位中通貨膨脹避險基金佔有最大的成份，債券持有部位中通貨膨脹避險基金與狀態變數避險基金佔有最大的成份。 關鍵字：確定提撥、薪資的不確定性、通貨膨脹、隨機控制、動態規劃 / In this study, we investigate the portfolio selection problem in order to hedge the labor income and inflation uncertainties for defined contribution (DC) pension schemes. First, we extend the previous work of Battocchio and Menoncin (2004) that allowed the state variables (i.e., the risks from the financial market) and a set of stochastic processes to describe the inflation, labor income and expense uncertainties. A five-fund separation theorem is derived to characterize the optimal investment strategy for DC pension plans to hedge the labor income and the inflation risks. Second, by solving the Hamiltonian equation in the three-asset framework, we show that the optimal portfolio consists of five components: the myopic market portfolio, the hedge portfolio for the state variables, the hedge portfolio for the inflation risk, the hedge portfolio for the labor income uncertainty and the riskless asset. Then we explicitly solve the optimal portfolio problem. Finally, the numerical results indicate that the inflation hedge portfolio comprises the overwhelming proportion of stock holdings in the optimal portfolios. In addition, the inflation hedge portfolio and the state variable hedge portfolio constitute the overwhelming proportions of bond holdings. Keywords: defined contribution; salary uncertainty; inflation; stochastic control; dynamic programming.

本研究探討於連續時間下，跨國投資者於匯率可預測下之最適投資決策問題。我們假設隨著時間改變，利用可預測之資訊動態修正投資決策。首先我們假設匯率可經由利率過程預測，探討相對風險趨避(CRRA)之投資經理人於跨國投資時之避險需求。研究方法是結合Cox與Huang (1989)之平賭方法與Lioui與Poncet (2003)於跨國投資所建構之財務模型。本研究歸納學習效果會影響匯率期望報酬，利用利率資訊會修正匯率過程之風險市場價值。最適投資決策因此受到調整因子之影響。因此投資人必須依照過濾進來的財務訊息(利率對匯率的改變)動態的調控持有之投資部位。最後，理論結果顯示投資部位必須針對可預測性下匯率避險效果作調整。 / In this study, we explore the effects of uncertainty about the exchange rate predictability on international portfolio choice in a continuous time setting. Uncertainty regarding to the predictive relation affects the optimal portfolio choice through dynamic learning, and leads to a state-dependent relation between the optimal portfolio choice and the investment time horizon. First we investigate the hedge demands in international portfolio management for constant relative risk averse investors where the exchange rate can be predicted by the change of interest rate. Then our approach is implemented through the use of the martingale methodology developed by Cox and Huang (1989) as proposed in the work of Lioui and Poncet (2003). Since the learning processes influence the premium of exchange rate movements, the crucial changes lie in the difference of market price of risk of the interest rate movements to the updated exchange rates. The constructed optimal investment strategy is influenced by the adjusted factors. Hence the investors should dynamically rebalance their holding portfolio according to the filtering mechanism. Finally, the theoretical results show that the adjustment for the optimal weights are required to reflect the prediction effects in hedging the exchange rate risks.

在觀察性研究或非隨機試驗研究中，欲探討因果效應時，研究者需要重新對觀察性研究進行設計，設計目的在於重新建立一個隨機指派受試者的機制，使其得以近似一個隨機試驗研究，這樣的研究一般稱為「類隨機試驗研究」(quasi-randomized-experiments)。 傾向分數分析即為一種設計觀察性研究的方法，在不牽涉到反應變數結果之下進行設計。本文於一個病例對照研究(case-control study)中使用傾向分數進行配對接著再進一步估計處理效果，傾向分數配對是可降低觀察性研究中的選擇性偏誤的方法，透過配對可減少實驗組與對照組間的系統性差異，使研究群體在所觀察到的控制變數分配達到相似，進而得到處理效果(treatment effect)的不偏估計，為近年廣受流行病學、經濟學以及社會學領域使用的方法之一。傾向分數本身為一個條件機率，定義為研究受試者在其所觀察到的控制變數之下，接受某處理或被指派至某特定群體的機率，估計傾向分數最常見的方法為羅吉斯迴歸。 此外，自1970年代起，配對方法(matching method)開始被使用來選取合適的實驗組與對照組並進行兩群體的比較，其中「確切配對」屬於最常使用的配對方法，過去文獻中經常可見各種配對方法的結合使用，因此，本文電腦模擬研究部份，欲比較四種情境之下「傾向分數配對」與「確切配對」結合使用的效果，分別以偏誤降低比例、信賴區間覆蓋率、均方誤衡量兩種配對方法結合使用的適合情境。結果顯示若對「與處理指派中度相關的變數」且「與反應變數高度相關的變數」，其效果最為明顯。根據結果，我們總結認為「確切配對與傾向分數配對合併使用」確實會有較好的表現，但表現的好壞也取決於確切配對的變數。實證研究部份，探討家庭結構對青少年偏差行為之影響，欲了解來自非完整家庭之青少年是否較來自完整家庭之青少年更有容易出現偏差行為。 / In observational or nonrandomized studies, treatments are not randomly assigned so that baseline differences between treated and control groups are typically observed. Without properly executed, the differences would bias the treatment effect estimates. There has been a long history of using matching to eliminate confounder bias, and inferences are made based on the matched observations. The theoretical basis for matching has been developed since 1970, and among those matching methods commonly in use, the exact matching is probably the most popular one. On the other hand, introduced by Rosenbuam and Rubin in 1983, propensity scores, the conditional probability of being exposed or treated given the observed covariates, has been a welcome alternative used to adjust for baseline differences between study groups of late. Instead of matching a treated with an untreated subject by their covariates, subjects in both treated and control groups are matched by their propensity scores. In this study, we explore the benefits of using propensity score matching together with the exact matching for adjusting for baseline differences through Monte Carlo simulations. An empirical study is also be provided for illustration.