同時比較多個樣本間的差異，可用ANOVA來檢定，但ANOVA只能得到樣本間有差異的訊息，無法明確指出是哪些樣本間有差異，需要使用多重比較找出樣本間的差異。本文主要探討相關的離散型資料的多重比較，以型I誤差與檢定力兩指標找出最適的多重比較法。本文依序探討獨立的連續型資料、相關的連續型資料、獨立的離散型資料、相關的離散型資料，並針對相關型的資料提出修正法。綜合型I誤差與檢定力兩指標來看，在樣本間的平均差異小時，Shaffer’s first procedure Test (1986)、Procedure 4 by Bergmann and Hommel (1988)為兩兩比較下較佳的修正法，Hochberg Test (1988)為多對ㄧ比較下較佳的修正法；樣本間平均差異大時，Bonferroni 為兩兩比較下較佳的修正法，Hochberg (1988)、Simes (1986)為多對ㄧ比較下較佳的修正法。 / Analysis of variance (ANOVA) is usually applied to check whether there are differences among more than two treatments. However, even there are differences, multiple comparison procedures are still needed to determine which pair(s) of treatments are different. In this study, we use simulation to compare the frequently used multiple comparison procedures, including many-to-one and pair-wise, and type-I error and power are used to measure the performance of procedures. Two types of data were considered, independently and correlated distributed data. If the differences among treatments are small, Shaffer’s first procedure test (1986) and Procedure 4 by Bergmann and Hommel (1988) are the best in pair-wise case, and Hochberg test (1988) is the best in many-to-one case. If the differences among treatments are large, the Bonferroni procedure is the best in pair-wise case, and the procedures by Hochberg (1988) and Simes (1986) are the best in many-to-one case.

本文將模糊理論的觀念，應用在估計、檢定及時間數列分析上。研究重點包括離散型及連續型模糊樣本的定義與度量，模糊參數的最佳估計，模糊排序方法應用於無母數檢定，模糊相似度的定義、性質，以及如何將其應用於辨識不同時間數列間的落差l期相似程度等。我們首先將常見的模糊資料分為離散型及連續型，並針對不同類型的資料，給定對應的模糊平均數、模糊變異數等模糊參數的概念與一些重要性質。接著我們提出幾種估計方法，針對不同的模糊參數進行最佳估計並提出可行的評判準則。進一步地，我們將模糊排序方法應用於無母數檢定推論。最後我們提出模糊相似度的定義與度量。經由系統性的模擬與分析，我們建立兩時間數列間模糊相似度演算法則。實證分析方面，我們利用提出的方法對台灣的股價加權指數、個股股價進行估計及檢定；同時，針對台灣歷年GDP、民間消費、毛投資間的相似性進行偵測，以驗證我們提出的模糊參數估計、模糊無母數檢定及模糊相似度演算法的效率性與實用性。 / In this paper, we apply fuzzy theory in estimation, nonparametric test, and time series analysis. Our focus is on: How to define and measure the discrete type fuzzy data and continuous one? How to find the optimal estimators for fuzzy parameters? How to apply fuzzy ranking methods in nonparametric test when the data is vague? How to define and find the degree of fuzzy similarity between two time series? First, fuzzy data is classified according to its type, discrete or continuous. Then we give some definitions and properties on fuzzy mean, fuzzy variance for different type of fuzzy data. Next, we proposed some estimating methods and evaluation rules. Moreover we apply fuzzy ranking methods in nonparametric test, such as Sign test, Wilcoxon signed rank test, Wilcoxon rank sum test, and so on. Finally, we suggest the definitions as well as the algorithm for computing the degree of fuzzy similarity between two time series. We also give some simulate and empirical examples to illustrate the techniques and to analyze fuzzy data. Results show that fuzzy statistics with soft computing are more realistic and reasonable for the social science research.

過去研究指出，公司管理當局可藉由即時揭露更多攸關資訊以降低市場之資訊不對稱，而管理當局所發布之財務預測亦為揭露資訊之一種。我國證管會於民國八十年五月起正式實施強制性財務預測制度，影響資本市場甚鉅。本研究旨在探討管理當局所發布之財務預測對資本市場資訊不對稱之影響，進而推論我國強制性財務預測制度對於降低資本市場之資訊不對稱是否有其功效。 本研究採用股票交易量、股價變異性及市場深度作為資訊不對稱之代理變數，實證結果顯示： 1.在強制性財務預測制度實施前，自願發布財務預測之公司於預測發布後，其資訊不對稱顯著較預測發布前降低；然與未發布財務預測公司相較之結果卻顯示，以股票交易量為資訊不對稱之代理變數時，發布財測公司於預測發布後之資訊不對稱反而顯著較未發布財測者為高。 2.在強制性財務預測制度實施後，強制或自願發布財務預測之公司於預測發布後，其資訊不對稱程度仍顯著較其發布前降低；而以股價變異性為資訊不對稱代理變數之結果亦顯示，發布財測公司於預測發布後之資訊不對稱顯著較未發布公司為低。 3.以市場深度為資訊不對稱代理變數之結果顯示，在強制性財務預測制度實施後，發布強制性財務預測之公司，其資訊不對稱於預測發布後顯著降低；此外，與發布自願性財務預測公司相較之結果顯示，發布強制性財測公司於預測發布後，其資訊不對稱程度不顯著高於發布自願性財務預測者。 / A firm can increase levels of disclosure to lower the information asymmetry. Financial forecast released by managers is also one of information about corporation. Our country began to implement the mandatory financial forecast regulations since May, 1991. This study mainly investigates the association between financial forecast released by companies and the mandatory financial forecast regulations. Furthermore , it also investigates that if the regulations effectively mitigate information asymmetry. This study uses trading volume, price volatility, and market depth as proxies for the information asymmetry. The empirical results show that: 1.Before May, 1991, corporations with voluntary forecast significantly mitigated the information asymmetry after the forecast released. But the information asymmetry (use trading volume as a proxy) of corporations after forecast released was not significantly lower than corporations without forecast. 2.After May, 1991, corporations with mandatory or voluntary forecast also significantly mitigated the information asymmetry after the forecast released. And the information asymmetry (use price volatility as a proxy) of corporations after forecast released was significantly lower than corporations without forecast. 3.After May, 1991, corporations with mandatory forecast significantly mitigated the information asymmetry (use market depth as a proxy) after the forecast released. And the information asymmetry of corporations after mandatory forecast released was not significantly higher than corporations with voluntary forecast.

本研究主要是探討選擇權成交量與期貨價格變動之關係，檢定Schlag and Stoll (2005) 的三個假說：完全效率市場假說（perfect market hypothesis）、資訊假說（information hypothesis）及流動性假說（liquidity hypothesis）。採用2005年一整年期貨與選擇權之日內資料，將同一日內樣本資料區分成盤前、盤中與盤後三個交易時段，另外亦將同一週內區分成不同交易日（週一至週五），藉由一般化自我相關條件異質變異模型(GARCH)，得出下列三個實證結果： / 1. 日內未區分交易時段下，拒絕完全效率市場假說及資訊假說，淨買權或是正向選擇權成交量對於期貨價格有正向之影響，淨賣權或是負向選擇權成交量對於期貨價格則有負向之影響，且選擇權成交量對於期貨價格變動之同步價格效果在接下來的六分鐘內反轉。無法拒絕流動性假說。 / 2. 日內區分交易時段、但週內不區分不同交易日下，無論盤前、盤中及盤後，皆拒絕完全效率市場假說及資訊假說，而不拒絕流動性假說。 / 3. 日內區分交易時段、且週內區分不同交易日下，週內各交易日盤前與盤中時段，皆拒絕完全效率市場及資訊假說，而無法拒絕流動性假說；而週內盤後時段結果則相對複雜，值得進一步討論。

信用風險是金融機構最關切的風險之一，信用評等則是具有公信力的評等公司對企業債信良窳的客觀評估。本研究目的即在於探討在效率市場的假設前提下，股票價格所蘊含的信用風險與信用評等間之相互關係。我們以接受信用評等的上市櫃公司為研究對象，利用KMV模型求解出違約距離(Distance-to-Default, DD)，再使用Kalman Filter粹取出符合公司之市場資訊(Adjusted-DD)替代股票價格，本研究分成兩部分討論，第一部份以順序羅吉斯模型(Ordered Logit Model)及羅吉斯模型(Logit Model)探討股價變動是否能領先告知未來公司信用評等的變化情形。第二部分則利用一般化自我相關條件異質變異模型(GARCH Model)觀察信用評等變動是否為市場帶來新的資訊。 第一部分實證結果發現：當長期信評調降時，電子、通訊相關產業及金融業之市場資訊(Adjusted-DD)的變動與長期信用評等為負相關，而長期信用評等調升時，仍得到負向關係，短期信用評等之部分得到當信用評等調降時，電子與通訊相關產業市場資訊變動有負向關係，與長期信用評等得到一致性之關係，但金融業則顯示市場資訊與信用評等調降為正相關，而傳統產業顯示短期信用評等調升與市場資訊呈現負向關係。 第二部分實證結果：與大多學者之研究相符，當信用評等調降時股票價格有負的異常報酬，而本研究更進一步發現當信用評等調升時，股票價格同樣隨著上漲且有顯著結果，兩者具有對稱關係。故信用評等改變時能夠為市場帶來新的資訊，可視為投資的重要參考指標之一。但以股票價格所蘊含的信用風險與信用評等間的關係卻仍無法得到應證。

日前由於金融海嘯的發生，導致全球金融市場對於結構債此種收益較高的商品存在眾多疑點，然而究其原因乃投資人無法正確地了解到自身的風險屬性，盲目地追求高收益率，反而造成投資結果不盡理想。本文將應用模型來推導商品的理論價格，並深入分析結構債可能帶來的風險與報酬。 本文兩個案商品之連動標的分別為利率與股權。第一個個案商品為英國勞埃德TSB銀行所發行之十年期累積計息利率連動債，在評價上將採用LIBOR市場模型，利用市場上既有的資料校準模型所需參數與期初遠期利率；此外由於本商品具有提前贖回特性，故將應用最小平方法蒙地卡羅來找出該商品發行之期初價格，並分別就發行機構探討其避險策略及投資人應注意的風險作深入分析。 第二個個案商品為摩根大通公司所發行之附部分保本之股權連結式自動贖回債券，利用對數常態股價模型及蒙地卡羅法評價出其理論價格，並針對發行機構可能面對之風險與避險策略作完整探討，最後就投資人之報酬及風險層面作詳盡地剖析。

行人偵測為物件偵測領域中一個極具挑戰性的議題。其主要問題在於人體姿勢以及衣著服飾的多變性，加之以光源照射狀況迥異，大幅增加了辨識的困難度。吾人在本論文中提出利用共變異矩陣描述子及結合單純貝氏分類器與級聯支持向量機的線上學習辨識器，以增進行人辨識之正確率與重現率。 實驗結果顯示，本論文所提出之線上學習策略在某些辨識狀況較差之資料集中能有效提升正確率與重現率達百分之十四。此外，即便於相同之初始訓練條件下，在USC Pedestrian Detection Test Set、 INRIA Person dataset 及 Penn-Fudan Database for Pedestrian Detection and Segmentation三個資料集中，本研究之正確率與重現率亦較HOG搭配AdaBoost之行人辨識方式為優。 / Pedestrian detection is an important yet challenging problem in object classification due to flexible body pose, loose clothing and ever-changing illumination. In this thesis, we employ covariance feature and propose an on-line learning classifier which combines naïve Bayes classifier and cascade support vector machine (SVM) to improve the precision and recall rate of pedestrian detection in a still image. Experimental results show that our on-line learning strategy can improve precision and recall rate about 14% in some difficult situations. Furthermore, even under the same initial training condition, our method outperforms HOG + AdaBoost in USC Pedestrian Detection Test Set, INRIA Person dataset and Penn-Fudan Database for Pedestrian Detection and Segmentation.

本論文探討法語借詞進入國語音韻系統時所採取的音韻調整策略，並從優選理論和語料庫的角度來分析借詞現象，主軸為法語子音和鼻母音的調整模式。 本研究建置一個法語借詞語料庫，並統計音段保留、替代、或刪除的趨向。研究發現，音段保留與調整為主要的轉譯策略，唯有韻尾子音[ʁ]傾向於刪除。數據分析顯示法語借詞中的語言變異(variation)極為普遍，此現象說明國語中的法語借詞尚未完全詞彙化 (lexicalization)，而是仍在持續進行。因此本論文採用ROE模型 (rank-ordering model of EVAL, Coetzee 2006)來解釋法語借詞轉譯的變異現象。 法語音段的調整現象和變異情形可訴諸於制約的排序與互動。根據ROE模型，制約係以中界線(cut-off line)加以區隔，違反中界線以上的候選值會直接淘汰，而僅違反中界線以下的候選值皆可成為優選值。本研究發現，中界線以上的制約包含結構制約、信實制約、以及聯合制約(conjoined constraint);中界線以下的制約則多為信實制約。音韻表徵(phonological features)如[labial]與[nasal]在轉譯的過程中必須保留，而是否違反[voice]與[spread glottis]等表徵僅會造成語言變異。 制約的交錯排序、互動，以及中界線的位置對於法語借詞的現象提出了解釋。借詞音韻並非自成一個系統，而是反映了借入語和普遍語法皆存在的制約。 / This thesis investigates the phonological adaptations of French loanwords in Mandarin. The focus is on the loanword adjustment of French consonants and nasal vowels. A loanword corpus is established. Based on the statistical analysis, the adaptation pattern of each foreign segment is provided. The thesis observes that preservation and adaptation are the major strategies in loanword adaptation. The only exception is [ʁ] in the coda position, which has segment deletion as the major substitute. Variant substitutes are common in French loanwords, which suggest that the loanword adaptation of French is an ongoing process in Mandarin. Optimality Theory (Prince and Smolensky 1993/2004) is the major framework in this thesis. To account for variations, Coetzee’s (2006) rank-ordering model of EVAL (ROE) is also used for analysis. A critical cut-off line divides the constraints into two parts: constraints above the cut-off line, and constraints below the cut-off line. Violations of the former are fatal, while violations of the latter lead to variations. The fatal constraints place restrictions on illicit segments, segment combinations, and prosodic preferences. These crucial constraints include markedness constraints, faithfulness constraints, or even the conjunction of the two. Constraints below the cut-off line are mostly IDENT constraints. Violations of IDENT features such as [voice] and [spread glottis] bring only variations. This is contrary to the violations of IDENT [nasal] and IDENT [labial], which result in ill-formedness. Based on the constraint ranking and the position of the cut-off line, the thesis shows that loanword phonology does not form an idiosyncratic grammar. Instead, it includes the constraints that are latent in L1 and reflects universal grammar.

Most indexes use a cap-weighted strategy as the asset allocation method. Many researchers suggest that although the cap-weighted strategy often serves as an appropriate surrogate to the market portfolio, it cannot consistently outperform other portfolio weighting strategies. The main reasons behind it would be related to the market movement and the underlying volatility. It is possible to narrow down the scope of the research by focusing on an industry index as was done in this research. The focus would be on the mining index of Indonesia. Comparisons will be done between the established index to other portfolio weighting schemes, namely the equally weighted portfolio and the minimum variance portfolio. The results of the research was that the index was not quite efficient, both on the returns and the Sharpe or information ratio aspect. There are many possible reasons behind this, but the most possible reasons would be that the stocks included do not contribute to diversification, over focus on the coal industry, lack of rebalancing or restructuring, in addition to the market itself. The implication of this research would be that stock exclusion might also contribute to risk minimization.

處理類別型的資料時，常由於變異數與平均數間具有函數關係，因此資料呈現出來的變異程度會比預期的變異程度來的大，這種現象就稱為資料具有額外變異。一般的分析方法是利用廣義線性模型先作估計，再對估計之標準誤做調整。本文中將探討處理額外變異的另外兩種方法—準概似估計和隨機效果模型，並分別利用紡織原料與毒物學研究之資料作為範例來比較此兩種方法與前者的異同。 / Overdispersion is a common phenomenon in practice when modelling categorical data, and the scaled Pearson chi-square is usually used to measure it. In this study, we examine two other methods—the quasi-likelihood and the random-effect models. In addition, two examples are provided for illustration.