環境保護議題在今日已成為人類經濟活動中，除了「全球化」以外另一個引人重視的議題，而為達到環境保護，以維持共生環境之永續發展之目的，減少環境資源消耗與廢棄物的排放（節能省碳），可以說是效用最大的方法。歐盟在2006年7月1日RoHS法令開始施行後，在可預見的短期時間內，注重或從事環境保護相關的企業，將格外具有投資的價值，而在另一方面，隨著新興市場經濟的急速發展，對於資源的需求殷切，使得國際原油價格居高不下，並且可能持續漲升，所以從事節能相關的研究與應用，不僅是未來的趨勢，亦是企業未來的商機所在。因此，本研究選定具有環保及節能雙重利基的發光二極體（light-emitting diode, LED）晶粒產業進行研究與分析，目的在於提供企業經營者對於未來經營策略方向之參考，以及提供投資人採用價值投資法進行投資之考量依據，而研究結果對於後續研究者，亦將具有相當程度之參考價值。 本研究項目內容包括研究動機與目的及研究範圍與限制、文獻回顧與探討、產業整體分析、評價模式說明、實證研究結果與分析及結論與建議等六大項。 本計畫研究方法採用三階段現金流量折現模式。首先，對LED產業進行整體分析，然後就個案公司營運情況進行企業評價之前提分析，並將個案公司初步評價結果與股價歷史資料比較，以檢視評價結果之績效，且以產業未來可能發展景況評估個案公司股價合理區間；其次，將評價結果與目前市場股價作比較，並探討目前股價背後所隱含之變數值；最後利用敏感度分析來觀察個別關鍵評價因子對股價之影響程度。 在分別對銷售導向及盈餘導向DCF法採用不同權重進行綜合評價時，可得知晶電、璨圓、華上及泰谷公司，在銷售導向及盈餘導向DCF法評價結果之權重值分別為60%-40%、70%-30%、70%-30%及70%-30%時，有相對較低之年平均Theil’s U值0.272516、0.190505、0.138237及0.135523，即在該權重值下各個案公司之綜合評價結果可獲得相對較佳之評價績效。依據上述較佳評價績效之權重值，以2007年12月31日為評價日，進行晶電、璨圓、華上及泰谷公司之綜合評價，其合理的股價區間分別為每股61.07~134.88元、16.7~32.05元、9.87~31.19元及10.86~32.69元。 在影響股價變動的關鍵因子判斷上，由分析結果可知晶電公司盈餘成長率、總投資率及銷售成長率等因子每變動1%，對於股價之影響程度（包含正向或負向變動）皆大於1%，顯示此三項因子應為影響晶電公司股價變動之關鍵因子；對於璨圓公司而言，影響股價變動之關鍵因子則為銷售成長率及盈餘成長率；而影響華上公司股價變動之關鍵因子則僅有總投資率；至於影響泰谷公司股價變動之關鍵因子則包括總投資率、盈餘成長率、邊際利潤率及銷售成長率等4項。

本文主要藉由逐次改變二次擔保債權憑證的內層分券金額佔資產池發行金額比例、內層分券下層信用保護金額佔資產池金額比例、資產池參考標的間違約相關性、到期期限、及違約回收率等五項影響二次擔保債權憑證損失發生機率的風險因子，結合蒙地卡羅模擬法及關聯結構法模擬交易架構中內層、外層分券不同損失率的發生機率，並利用彈性分析，衡量二次擔保債券憑證在每單位風險因子變動下，內層及外層分券的損失發生機率。 研究結果顯示，相同的風險因子對於內層與外層分券的損失發生機率的影響效果並不相同，此一現象有別於一般認為風險因子對內、外層分券損失發生機率影響效果相同的看法。此外，依據分券損失發生機率對每單位風險因子變化的彈性敏感性分析，分券損失發生機率受風險因子的影響可分為: 彈性為正且數值逐漸增加、彈性為正且逐漸下降、彈性為負且數值 (絕對值) 逐漸下降、及彈性為負且數值 (絕對值) 逐漸增加四類。外層夾層分券的損失發生機率對內層分券厚度占資產池金額比例的彈性為負，其數值 (絕對值) 隨著內層分券厚度占資產池金額比例的增加而下降。外層夾層分券的損失發生機率對內層分券下層信用保護金額佔資產池金額比例的彈性、及外層夾層分券的損失發生機率對參考標的違約回收率的彈性為負，且數值 (絕對值) 隨著下層信用保護比例及回收率的增加而上升。外層夾層分券的損失發生機率對參考標的違約相關係數的彈性為正，其數值隨著相關係數的增加而下降；外層夾層分券的損失發生機率對參考標的之到期期限的彈性為正，其數值隨著到期期限的增加而上升。 / The researchers of this study combined Monte Carlo simulation approach and copula method to change the following five risk factors: the thickness of inner CDOs tranche on CDO-squared, the subordination in master CDOs tranche, the correlation of reference entities, the maturity of reference entities, and the recovery rate of reference entities, with a purpose of simulating the loss possibility of CDOs-squared. Besides, by elasticity analysis, the researchers measured the change of loss rate according to the change of each risk factor per unit. The result of the study shows that the same risk factor has different influence on the loss rate of inner and master tranche of CDOs squared, which mismatches the general belief that the same risk factor has the same effect on the loss rate of inner and master CDOs tranche. In addition, according to the tranche loss possibility elasticity analysis to the risk factors, this research reveals that four categories can be made due to the effect which risk factors have on loss rate : positive and increasing elasticity, positive and decreasing elasticity, negative and increasing elasticity, and negative decreasing elasticity. We found that for the master mezzanine tranche: the elasticity of tranche loss possibility to the thickness of inner CDOs tranche of CDO-squared is negative and will decrease with the increasing thickness of inner CDOs tranche. The elasticity of tranche loss possibility to subordination in inner CDOs tranche and the elasticity of tranche loss possibility to the recovery rate of reference entities are both negative and will increase with the increasing subordination of inner CDOs tranche and the recovery rate of reference entities. The elasticity of the loss rate possibilities to the correlation of reference entities default is positive and will decrease with the increasing correlation of reference entities. The elasticity of loss possibilities to the maturity of reference entities is positive and will increase with the increasing maturity.

一、研究動機：設備更新是經常面臨的問題。設備更新分析已發展出多種模式，可是 廠商未能充分使用，乃引發本研究對其中問題再做探討。 二、研究架構：(一)列舉設備更新面臨的問題。(二)探討各種模式是否解決前項問題 并做小幅度修正。(三)比較與選用模式。 三、研究重點：(一)綜合整理各種更新模式并比較應用。(二)時間幅度及不確定性對 各種模式及決策的影響。(三)設備更新計劃之擬定。(四)實證研究。 四、研究方法：(一)由文獻收集各種更新模式。(二)應用動態規劃法、隨機法、模擬 法等新式方法。(三)各種模式做敏感度分析,以供決策參考。

Stacy (1962)首先提出廣義伽瑪分配 (generalized gamma distribution)，此分布被廣泛應用於存活分析 (survival analysis) 以及可靠度 (reliability) 中壽命時間的資料描述。事實上，像是指數分配 (exponential distribution)、韋伯分配 (Weibull distribution) 以及伽瑪分配 (gamma distribution) 都是廣義伽瑪分配的一個特例。 Bologna (1987)提出一個特殊的雙變量廣義伽瑪分配 (bivariate generalized gamma distribution) 可以經由雙變量常態分配 (bivariate normal distribution) 所推得。我們根據他的想法，提出多變量廣義伽瑪分配可以經由多變量常態分配所推得。在過去的研究中，學者們做了許多有關雙變量伽瑪分配。當我們提到雙變量常態分配，由於其分配的型式為唯一的，所以沒人任何人對其分配的型式有疑問。然而，雙變量伽瑪分配卻有很多不同的型式。 在這篇論文中的架構如下。在第二章中，我們介紹並討論雙變量廣義伽瑪分配可以經由雙變量常態分配所推得，接著推導參數估計以及介紹模擬的程序。在第三章中，我們介紹一些對稱以及非對稱的雙變量伽瑪分配，接著拓展到雙變量廣義伽瑪分配，有關參數的估計以及模擬結果也將在此章中討論。在第三章最後，我們建構參數的敏感度分析 (sensitivity analysis)。最後，在第四章中，我們陳述結論以及未來研究方向。 / The generalized gamma distribution was introduced by Stacy (1962). This distribution is useful to describe lifetime data when conducting survival analysis and reliability. In fact, it includes the widely used exponential, Weibull, and gamma distributions as special cases. Bologna (1987) showed that a special bivariate genenralized gamma distribution can be derived from a bivariate normal distribution. Follow his idea, we show that a multivariate generalized gamma distribution can be derived from a multivariate normal distribution. In the past, researchers spend much time in working on a bivariate gamma distribution. When a bivariate normal distribution is mentioned, no one feels puzzled about its form, since it has only one form. However, there are various forms of bivariate gamma distributions. In this paper is as following. In Chapter 2, we introduce and discuss the bivariate generalized gamma distribution, then the multivariate generalized gamma distribution is derived. We also develop parameters estimation and simulation procedure. In Chapter 3, we introduce some symmetrical and asymmetrical bivariate gamma distributions, then they are extended to the bivariate generalized gamma distributions. Problems of parameters estimation and simulation results are also discussed in Chapter 3. Besides, sensitivity analyses of parameters estimation are conducted. Finally, we state conclusion and future work in Chapter 4.

Fong與Vasicek（1997）提出風險分析應考慮敏感度分析、風險值及壓力測試，才能完整揭露投資組合的風險狀況。其中風險值的計算，不僅考慮二階風險，並且利用三階動差進行偏態修正。本文除了以變異數-共變異數法、歷史模擬法及蒙地卡羅模擬法此三種方法計算風險值，並利用Fong與Vasicek（1997）偏態修正法及Cornish-Fisher偏峰態修正法來做偏態及峰態的修正。而後再利用概似比檢驗法、回溯測試百分比法及Z檢定法作為驗證風險值模型的評比工具。我們建議在95%及99%的信賴水準下，求算風險值可利用Cornish-Fisher所提出的方法修正偏態及峰態。 / Fong and Vasicek (1997) mentioned that risk analysis should include sensitivity analysis, value at risk (VaR) and stress testing, in order to capture portfolio risk. The calculation of VaR should not only consider the second moment but should also adjust the skewness using the third moment. In this article, we determine VaR by employing three methods, the variance covariance, the historical simulation and the Monte Carlo simulation methods. In addition, we also adjust VaR for the skewness and kurtosis using the methods developed by Fong and Vasicek (1997) and Cornish-Fisher. Then, the likelihood ratio test, back testing and the Z-test are used to verify the VaR model. Our final test results suggest that calculating VaR should be adjusted for the skewness and the kurtosis as shown by the method proposed by Cornish Fisher in the 95% and 99% confidence intervals.

本研究的主要目的是檢視倒傳導神經網路是否具有人類在分類學習上所呈現出來的學習效應 — 競爭學習、遮蔽效應與不相關線索的影響。在實驗中，我們採用兩種倒傳導神經網路，來測試激發函數是否會影響倒傳導神經網路的學習。此兩種倒傳導神經網路分別採用sigmoid激發函數與hyperbolic-tangent激發函數。實驗結果顯示，以sigmoid為激發函數與以hyperbolic-tangent為激發函數的倒傳導神經網路都具有這三個學習效應。還有，以sigmoid為激發函數的倒傳導神經網路所呈現出來的學習效應比以hyperbolic-tangent為激發函數的倒傳導神經網路來得顯著。本研究的次要目的在於瞭解有效性（使用性）與敏感度分析的數值是否有對應關係。實驗結果顯示，線索A與線索B的敏感度分析數值差異可以反映出線索A與線索B的有效性差異。然而，敏感度分析數值卻無法準確地顯示線索的有效性數值。 / The main objective of this research is to examine whether back propagation neural networks (BP) have the learning effects found in human category learning — competitive learning, overshadowing and the deleterious of an irrelevant cue. Two kinds of BP, BP with sigmoid activation function and BP with hyperbolic-tangent activation function, are investigated to see if the activation function will make BP behave differently. According to the results of our experiments, these three learning effects are demonstrated both in BP with sigmoid and BP with hyperbolic-tangent, but they seems more significant in BP with sigmoid than in BP with hyperbolic-tangent. The second objective of our research is to see if there is a correspondence between the validity (the utilization) and the value of sensitivity analysis, R. From the results of our experiments, we observe that the difference between values of sensitivity analysis with respect to Cue A and Cue B reflects the difference of the validities between Cue A and Cue B. However, the value of sensitivity analysis does not show exactly what validity a cue is.

本篇論文旨在分析神經網路學習績效，並提出一套學習演算法，結合倒傳遞網路(BP)與理解神經網路(RN)，命名為RNBP，這套學習演算法將與傳統的BP做比較，以兩個不同的財務金融領域的應用，一個是選擇權上Black-Scholes訂價模式的模擬，一個是金融交換上利率的預測。主要績效的評估準則是以學習的效率與模擬、預測的準確度為依據。 此外，本論文的另一個重點是提出一套對於神經網路系統進一步分析的方法與工具，敏感度分析(Sensitivity Analysis)與滯留區(Dead Region)分析，藉以瞭解神經網路系統是否具有效地良好學習或被一般化的能力，從神經網路的角度來說，這也是BP與RNBP的另一個績效比較標準。本研究的結果顯示RNBP在預測準確度上較BP為優良，但是在學習效率與預測能力的穩定性上並沒有呈現一致性的結論；此外，敏感度分析與滯留區分析的結果也幫助神經網路在應用領域上有更深入的瞭解。 在過去，神經網路的應用者往往忽略了進一步瞭解神經網路的重要性與可行性，本論文的貢獻在於藉由分析神經網路所學習的知識，幫助應用者進一步瞭解神經網路表達的訊息在應用領域上所隱含的實質意義。 / The study attempts to analyze the learning performance of neural networks in applications, and propose a new learning procedure for the layered feedforward neural network systems, named KNBP, which binds RN and BP learning algorithms. Two artificial neural networks, BP and KNBP, here are both applied to two financial fields, the simulation of Black-Scholes pricing model for the call options and the midrates forecasting in financial swaps. The explicit performance comparison between the two artificial neural network systems is mainly based on two criteria, which are learning efficiency and forecasting effectiveness. Then we propound a mathematical methodology of sensitivity analysis and the dead regions to deeply explore inside the network structures to see whether the models of ANNS are actually well trained or valid, and thus setup an alternative comparable criterion. The results from this study show that RNBP performs better than BP in forecasting effectiveness, but RNBP obtains neither a consistent learning efficiency in cases nor a stable forecasting ability. Furthermore, the sensitivity analysis and the dead region analysis provide a deeper view of the ANNs in the applied fields. In the past, most studies applying neural networks ignored the importance that it is feasible and advantageous to obtain more useful information via analyzing neural networks. The purpose of the research is to help further understanding to the information discovery resulted from neural networks in practical applications.

許多公司在發行公司債的時候，會給此公司債一個可提前贖回的特性，此種公司債稱為可贖回公司債(Callable Bond)，用來規避利率變動風險的金融商品也與我們熟知的利率交換不同，稱為可取消利率交換(Cancelable Swap)。其實可取消利率交換可以拆解成百慕達利率交換選擇權(Bermudan Swaption)加上利率交換，由於利率交換之評價較簡單也有市場一致的評價方法，因此百慕達利率交換選擇權便成為評價的重點。 評價的部分，由於百慕達式的商品有提前履約的特性，造成其封閉解不存在，因此需要利用其他的近似解或是數值方法來求它的價格。由於本文採用BGM(1997)的市場利率模型(Libor Market Model)，其高維度的性質導致數狀方法與有限差分法使用起來較無效率，因此本文選擇使用蒙地卡羅法做為評價的方法，同時利用Longstaff and Schwartz(2001)的最小平方蒙地卡羅法(Least Squares Monte Carlo Method)來解決提前履約的問題。 最後，本文將採用2種利率波動度假設與2種不同利率間相關係數的假設，共4種組合，在歐式利率交換選擇權的市場波動度下進行校準，使用校準出來的參數進行評價來得到4種價格。再進行商品的敏感度分析(Sensitivity Analysis)和風險值(Value at Risk)的計算。

本文使用Esscher轉換法推導狀態轉換模型、跳躍獨立風險下狀狀態轉換模型及跳躍相關風險下狀態轉換模型的選擇權定價公式。藉由1999年至2011年道瓊工業指數真實市場資料使用EM演算法估計模型參數並使用概似比檢定得到跳躍相關風險下狀態轉換模型最適合描述報酬率資料。接著進行敏感度分析得知，高波動狀態的機率、報酬率的整體波動度及跳躍頻率三者與買權呈現正相關。最後由市場驗證可知，跳躍相關風險下狀態轉換模型在價平及價外的定價誤差皆是最小，在價平的定價誤差則略高於跳躍獨立風險下狀態轉換模型。 / In this paper, we derive regime switching model, regime switching model with independent jump and regime switching model with dependent jump by Esscher transformation. We use the data from 1999 to 2011 Dow-Jones industrial average index market price to estimate the parameter by EM algorithm. Then we use likelihood ratio test to obtain that regime switching model with dependent jump is the best model to depict return data. Moreover, we do sensitivity analysis and find the result that the probability of the higher volatility state , the overall volatility of rate of return , and the jump frequency are positively correlated with call option value. Finally, we enhance the empirical value of regime switching model with dependent jump by means of calculating the price error.

Chen, Chang, Wen and Lin (2013)提出馬可夫調控跳躍過程模型(MMJDM)描述股價指數報酬率，布朗運動項、跳躍項之頻率與市場狀態有關。然而，利率並非常數，本論文以狀態轉換模型配適零息債劵之動態過程，提出狀態轉換下的利率與具跳躍風險的股票報酬之二維模型(MMJDMSI)，並以1999年至2013年的道瓊工業指數與S&P 500指數和同期間之一年期美國國庫劵價格為實證資料，採用EM演算法取得參數估計值。經由概似比檢定結果顯示無論道瓊工業指數還是S&P 500指數，狀態轉換下利率與跳躍風險之股票報酬二維模型更適合描述報酬率。接著，利用Esscher轉換法推導出各模型下的股價指數之歐式買權定價公式，再對MMJDMSI模型進行敏感度分析以評估模型參數發生變動時對於定價公式的影響。最後，以實證資料對各模型進行模型校準及計算隱含波動度，結果顯示MMJDMSI在價內及價外時定價誤差為最小或次小，且此模型亦能呈現出波動度微笑曲線之現象。 / To model asset return, Chen, Chang, Wen and Lin (2013) proposed Markov-Modulated Jump Diffusion Model (MMJDM) assuming that the Brownian motion term and jump frequency are all related to market states. In fact, the interest rate is not constant, Regime-Switching Model is taken to fit the process of the zero-coupon bond price, and a bivariate model for interest rate and stock index return with regime-switching and dependent jump risks (MMJDMSI) is proposed. The empirical data are Dow Jones Industrial Average and S&P 500 Index from 1999 to 2013, together with US 1-Year Treasury Bond over the same period. Model parameters are estimated by the Expectation-Maximization (EM) algorithm. The likelihood ratio test (LRT) is performed to compare nested models, and MMJDMSI is better than the others. Then, European call option pricing formula under each model is derived via Esscher transformation, and sensitivity analysis is conducted to evaluate changes resulted from different parameter values under the MMJDMSI pricing formula. Finally, model calibrations are performed and implied volatilities are computed under each model empirically. In cases of in-the-money and out-the-money, MMJDMSI has either the smallest or the second smallest pricing error. Also, the implied volatilities from MMJDMSI display a volatility smile curve.