本研究以使用者需求導向為設計原則，提出了建構物件與物件關聯資料庫工作量產生器的發展方式，以解決(1)一般績效評估標準規格因只針對特定應用領域，而無法周詳考慮到使用者的特定需求問題；(2)廠商所發展的工作量產生器則因為可能針對其產品的特定領域作設計與測試，而失去了公正性之問題。我們的目地就是希望能以此資料庫工作量產生器的設計方式使測試資料庫效能的工作量，對使用者而言更具一致性、代表性與正確性，如此測試之結果對使用者而言才能更具意義。 我們首先分析SQL3與ODMG標準，以此構成物件關聯與物件導向資料庫之一般化的工作量模式。然後，再著手以使用者需求導向為原則，設計工作量產生器的運作方式。最後，我們實際發展了工作量產生器的雛形系統，以讓使用者能真正依其需求建立工作量並實際作績效測試，並提供給欲建構物件與物件關聯資料庫工作量產生器的人員作參考。 / This thesis provides a method of establishing Object-Oriented and Object-Relational Database Workload Generator, and this method is designed from the perspective of the user's requirements for addressing two problems: (1) the benchmark standards are generally designed to specific domains, so they are unable think comprehensively users' special requirements. (2) the database workload generators that developed by database providers may lose justice, because they might be designed to their products' special domain. Our purpose is that hope this method can make the Workload more consistent, more representative and more accurate and the test result will be more meaningful to users. Firstly, we analyze the SQL3 and ODMG standards to compose the ORDBs' and OODBs' generalized Workload Models. Then, we design our Workload Generator form the perspective of the user's requirements. Finally, we develop the Workload Generator's prototype system, it can allow users to create Workload according to their actual requirements and really do the benchmark test, and furthermore can provide a reference resource to the people who want to develop ORDB and OODB Workload Generator.

本研究之主要目標在於建構一個能評估分散式關聯資料庫中之特色的需求導向績效評估方法。在過去的績效評估研究中，已經有許多人對於關聯式資料庫績效評估做了多方面的努力。但是，過去的關聯式資料庫資效評估方法如：Wisconsin、AS3AP、TPC系列的Benchmarks都有著一些限制及不足的地方。 過去的關聯式資料庫績效評估方法並無法完全的評估出分散式資料庫的特殊需求及其表現。所以本研究嘗試要建立出一個能專門適用於分散式資料庫導向的績效評估方法。為了要作出此績效評估方法，本研究採用了工作量模式的研究方法。先建出分散式資料庫績效評估的工作量模式，再以其來實作出績效評估方法。工作量模式分成三部分：資料模式、交易模式、控制模式。 / This thesis is intended to design a requirements-centric database benchmark, which can evaluate the general performance of the distributed relational database systems. In the past, there are many relational database benchmarks. But the relational database benchmarks like Wisconsin, AS3AP, TPC, TP1 have some constraints. In this study, we aim to design a general-purpose distributed database workload model and implement it. To design this benchmark, we need to build our workload model. The workload model consists of three components：data model, transaction model, control model. Each model has the requirement specification language to accommodate user's workloads.

政府利用公權力掌握之個人資訊包羅萬象，舉凡姓名、生日、身分證字號、家庭、教育、職業等。科技進步與網際網路發達，使原本散置各處之資料，可以迅速連結、複製、處理、利用；而為了增加行政效率與減少成本，機關透過網路提供公眾服務日益頻繁，藉由傳輸共用個人資料等情況已漸成常態。這些改變雖然對政府與民眾帶來利益，但是也伴隨許多挑戰，尤其當數機關必須共用資訊時，將使管理風險更添複雜與難度，一旦過程未加妥善管制，遭人竊取、竄改、滅失或洩露，不僅當事人隱私受損，也嚴重傷害政府威信。因此，凡持有個人資料的政府機關，均必須建立適當行政、技術與實體防護措施，以確保資料安全與隱密，避免任何可能危及資料真實之威脅與機會，而造成個人人格與公平之侵害。 　　隨著全球經濟相互連結以及網路普及，個人資料保護如今已是國際事務，這個趨勢顯現在愈來愈多的國家法律與跨國條款如OECD、歐盟、APEC等國際組織規範。而在先進國家中，美國與英國關於資訊隱私法制發展有其不同歷史背景，目前美國聯邦機關持有使用個人資料必須遵循的主要法規為隱私法、電腦比對與隱私保護法、電子化政府法、聯邦資訊安全管理法，以及預算管理局發布的相關指導方針；英國政府則必須遵守人權法與歐盟指令架構所制定的資料保護法，並且受獨立資訊官監督審核。此外，為了增加效率，減少錯誤、詐欺及降低個別系統維護成本，公務機關之間或不同層級政府所持有之個人資料流用有其必要性，故二國在資料傳輸實務上亦有特殊規定或作業規則。相較之下，我國2012年10月1日始施行的「個人資料保護法」對於公部門間傳輸個人資料之情形並無具體規定，機關內外監督機制亦付之闕如，使個人資料遭不當使用與揭露之風險提高。 為了保障個人資訊隱私權，同時使公務機關之間傳輸利用個人資訊得以增進公共服務而不違反當事人權益，本研究建議立法或決策者可參酌美國與英國法制經驗，明定法務部負責研擬詳細實施規則與程序以供各機關傳輸個人資料之遵循，減少機關資訊流用莫衷一是的情況；而為保證個人資訊受到適當保護，除了事先獲得當事人同意外，機關進行資料共用之前，應由專業小組審核，至於考慮採取的相關重要措施尚有：（1）建置由政策、程序、人力與設備資源所組成之個人資訊管理系統（PIMS），並使成為整體資訊管理基礎設施的一部分；（2）指派高階官員負責施行及維護安全控制事項；（3）教育訓練人員增加風險意識，塑造良好組織文化；（4）諮詢利害關係人，界定共用資料範圍、目的與法律依據；（5）實施隱私衝擊評估（PIA），指出對個人隱私的潛在威脅並分析風險減緩替代方案；（6）簽定正式書面契約，詳述相關權利與義務；（7）執行內外稽核，監督法規遵循情況，提升機關決策透明、誠信與責任。 關鍵詞：個人資料保護、隱私權、資訊隱私、資料傳輸、資料共用 / Governments have the power to hold a variety of personal information about individuals, such as the name, date of birth, I.D. Card number, family, education, and occupation. Due to advanced technology and the use of the Internet, personal data stored in different places can be connected, copied, processed, and used immediately. It is relatively common for government agencies to provide people with services online as well as transmit or share individual information to improve efficiency and reduce bureaucratic costs. These changes clearly deliver great benefits for governments and for the public, but they also bring new challenges. Specifically, managing risks around sharing information can sometimes become complicated and difficult when more than one agency is involved. If the government agency which keeps personal information cannot prevent it from being stolen, altered, damaged, destroyed or disclosed, it can seriously erode personal privacy and people’s trust in the government. Therefore, each agency that maintains personal data should establish appropriate administrative, technical, and physical safeguards to insure the security and confidentiality of data and to protect against any anticipated threats or hazards to the integrity which could result in substantial harm on personality and fairness to any individual . As the global economy has become more interconnected and the Internet ubiquitous, personal data protection is by now a truly international matter. The trend is fully demonstrated by the growing number of national laws, supranational provisions, and international regulations, such as the OECD, the EU or the APEC rules. Among those developed countries, both the U.S. and the U.K. have their historical contexts of developing legal framework for information privacy. The U.S. Federal agency use of personal information is governed primarily by the Privacy Act of 1974, the Computer Matching and Privacy Protection Act of 1988, the E-Government Act of 2002 , the Federal Information Security Management Act of 2002, and related guidance periodically issued by OMB. The U.K. government has to comply with the Human Rights Act and the Data Protection Act of 1998 which implemented Directive 95/46/EC. Its use of individual data is overseen and audited by the independent Information Commissioner. Further, because interagency data sharing is necessary to make government more efficient by reducing the error, fraud, and costs associated with maintaining a segregated system, both countries have made specific rules or code of practice for handling the transmission of information among different agencies and levels of government. By contrast, Taiwan Personal Information Protection Act of 2010 which finally came into force on 1 October 2012 contains no detailed and clear provisions for data transmitted between government agencies. Moreover, there are also no internal or external oversight of data sharing practices in the public sector. These problems will increase the risk of inappropriate use and disclosure of personal data. To protect individual information privacy rights and ensure that government agencies can enhance public services by data sharing without unreasonably impinging on data subjects’ interests, I recommend that law makers draw on legal experiences of the U.S. and the U.K., and specify that the Ministry of Justice has a statutory duty to prescribe detailed regulations and procedures for interagency data transmission. This could remove the fog of confusion about the circumstances in which personal information may be shared. Also, besides obtaining the prior consent of the data subject and conducting auditing by a professional task force before implementing interagency data sharing program, some important measures as follows should be taken: (1) Establish a Personal Information Management System which is composed of the policies, procedures, human, and machine resources to make it as part of an overall information management infrastructure; (2) Appoint accountable senior officials to undertake and maintain the implementation of security controls; (3) Educate and train personnel to raise risk awareness and create a good organizational culture; (4) Consult interested parties and define the scope, objective, and legal basis for data sharing; (5) Conduct privacy impact assessments to identify potential threats to individual privacy and analyze risk mitigation alternatives; (6) Establish a formal written agreement to clarify mutual rights and obligations; (7) Enforce internal as well as external auditing to monitor their compliance with data protection regulations and promote transparency, integrity and accountability of agency decisions. Key Words: personal data protection, privacy rights, information privacy, data transmission, data sharing

數年來台灣金控公司已如雨後春筍般冒出來，在金控公司底下含有產險公司、銀行、證券以及人壽公司等許多金融相關公司，因此，原本各自擺放於各子公司的資料庫可以通通整合在一起，當高階主管想提出決策時可利用資料庫進行資料採礦，以獲取有用的資訊。然而資料採礦的效果再怎麼神奇，也必須先有一個好的、完整的資料庫供使用，如果資料品質太差或者資料內容與研究目標無關，這是無法達成完美的資料採礦工作。 透過抽樣調查與函數映射的方法使得資料庫得以加值，因此當有目標資料庫與輔助資料庫時，可以利用函數映射方法使資料庫整合為一個大資料庫，再將資料庫中遺失值或稀少值作插補得到增值後的資料庫。在此給予這個整個流程一個名詞 ”Data SPA(Data Systematic Purifying Analysis)”，即「資料純化」。在本研究中，主要就是針對純化完成的資料進行結構的確認，確認經過這些過程之後的資料是效用且正確的。在本研究採用了橫向評估、縱向評估與全面性評估三種方法來檢驗資料。 資料純化後的資料經過三項評估後，可以發現資料以每個變數或者每筆觀察樣本的角度去查驗資料時，資料的表現並不理想，但是，資料的整體性卻是相當不錯。雖然以橫向評估和縱向評估來看，資料純化後的資料無法與原本完整的資料完全一致，但是透過資料純化的過程，資料得以插補且欄位得以擴增，這樣使得資料的資訊量增加，所以，資料純化確實有其效果，因為資訊量的增加對於要進行資料採礦的資料庫是一大助益。 / For the past few years, Taiwan has experienced a tremendous growth in its financial industry namely in banks, life and property insurances, brokerages and security firms. Needless to say the need to store the data produced in this industry has become an important and a primary task to accomplish. Originally, firms store the data in their own database. With the progressive development of data management, the data now can be combined and stored into one large database that allows the users an easy access for data retrieval. However, if the quality of the data is questionable, then the existence of database would not provide much insightful information to the users. To tackle the fore mentioned problem, this research uses functional mapping combining the goal and auxiliary database and then imputes the missing data or the rare data from the combined database. This whole process is called Data Systematic Purifying Analysis (Data SPA). The purpose of this research is to evaluate whether there is any improvement of the structure of the data when the data has gone through the process of systematic purifying analysis. Generally the resulting data should be within good quality and useful. After the assessments of the data structure, the behavior of the data with respect to their added variables and observations is unsatisfactory. However the manifestation of the data as a whole has seen an improvement. The modified database through Data SPA has augmented the database making it more efficient to the usage of data mining techniques.

縮減資料維度一直是多變量資料分析的主要技術, 運用投影追蹤法(Projection Purs -uit) 將高維度多變量資料投影至低維度空間, 亦是一種縮減維度的方法, 將更有助 於了解資料結構特性。 投影追蹤法最重要的部份是投影指數(Projection Index)的定義, 投影指數指出了使 用者的興趣與意圖, 并針對不同問題定義不同的投影指數, 也因此許多多變量分析技 術亦可視為投影追蹤法的特例, 例如: 將投影指數定義為投影后資料的變異數, 則推 導出主成分分析; 定義為一分類規則的錯誤率, 則可推導出線性判別分析……等等。 在探測性資料分析方面, 亦可定義出適當的投影指數以戡察資料的叢聚及非線性結構 等特性。定義投影指數之后, 則必須選取適當的數值分析方法(Numerical Method), 并輔以電腦功能來施行。除了縮減資料維度外, 投影追蹤法亦已於無毋數回歸分析(N -onparametric Regression Analysis)及密度估計(Density Estimation)的應用上大 有斬獲。 到目前為止, 投影追蹤法在多變量資料分析上, 仍為一新發展的技術, 在國內尚未見 有關此一研究範疇的報告及著作, 本文將就投影追蹤法這項新技術的起源、發展、沿 革、理論與應用做一廣泛而密集的探討。

本論文共一冊，約三萬字，全文共五章，茲表列如后： 第一章：緒論 第一節：研究動機與目的 第二節：基本假設與限制 第二章：基本定義與名詞 第一節：基本模型 第二節：常用的抽樣設計與估計式 第三章：最強的抽樣策略之探討 第一節：在超母體模型之下的最佳抽樣策略。 第二節：模型錯誤之處理情形 第三節：一些改進策略 第四章：實證研究 第五章：結論 本文所探討的主題是在一個具有超母體模型的有限母體之下估計母體總和，主要目的 為尋求一最佳抽樣策略，使得在所訂之條件下較其他抽樣策略均為佳，但上法是模型 指定沒有發生錯誤時的情形，但實際上模型的指定常發生錯誤，因此吾人有興趣找出 一最強抽樣策略使得模型指定錯誤時，該抽樣策略亦為最佳，最後再探討一些改進的 策略及實證研究。而以一些重要結論做總結。

近年來，資訊科技的進步，改變了人類溝通的方式、知識的管理和傳承、資訊的散播和儲存，對人類社會產生革命性的影響。檔案館開始面對大量的電子文書，並將原有的珍貴館藏轉換為數位物件典藏。然而，看似穩當的數位典藏方式，即使建立了適當的儲存環境，做好各種安全控管，未必能保證數位物件可供長期使用。為了長久保存數位物件並提供使用，除了轉存、轉置、技術模擬等廣受認同的保存方法外，還需要保存後設資料的配合，國際間已有許多相關研究。本研究以RLG、NLA、Cedars、NEDLIB四發展機構所提出的保存後設資料項目為比較對象，得出核心項目，再配合中研院近史所檔案館的需求，建置適用的保存後設資料項目。 經過比較發現，RLG、NLA、Cedars、NEDLIB所提出的保存後設資料項目，因為定位及應用目的不同，在規模、項目及描述層次上有差異，但描述重點仍有某種程度的共識。其中，NLA詳細記載歷次更動相關事項，並針對不同型態數位檔訂立不同的描述項目的做法，值得參考。 考量需求與館藏數位檔特性，中研院近史所檔案館保存後設資料項目雖以四種保存後設資料核心項目為基礎，在建置時仍做了部分修正與更動。(1) 保存後設資料的功能以協助保存決策制定及執行為主，不考慮提供使用的部分，也儘量不與原有後設資料重複；(2) 描述層級訂在系列層級；(3) 更動史項目，綜合NLA及Cedars之優點。限於系統功能限制，保存後設資料仍無法自動產生，也暫時不與數位檔及原有後設資料連結。 根據研究結果，對近史所檔案館及其他典藏機構提出下列建議：(1)制定數位保存政策及計畫；(2) 測試保存方法及保存後設資料項目；(3)留意相關科技及標準的發展趨勢；(4) 保存後設資料應儘量以自動、同步方式產生；(5) 明確定義描述規範。此外，對於數位保存後設資料及保存方法應再深入了解，並尋求跨領域合作機會。

本研究之方向係探究建置大型中醫藥倉儲所需之後設資料(Metadata)，並透過此一初探，瞭解與描述其中所需之分析方法，本研究設想Zachman Framework為合適之資料倉儲開發方法，因而由5W1H面向來衍生該專業領域所需之概念。此類概念可再透過一分析程序，確立後設資料。 因時間之限制，本研究採用「本草備要」、「醫方集解」為範例文件進行相關分析，以減少中醫流派林立及中文本身不準確性帶來之問題，當然，本研究在其中亦力求在整體架構上維持其他中醫藥典籍之適用性。 為達成目標，本研究首先探討了至今中草藥資料庫、資料倉儲、電子超文件領域之發展，因而本研究決定將個別之中醫典籍視為「資料專櫃」，而將分類樹、Metadata描述性資料置於目次(catalog)的概念之下，這樣的做法有利於整合其他典籍及其後設資料於大型資料倉儲中。 首先，本研究由重要中醫藥典籍導出基礎性中草藥概念與名詞，其後透過典藏面及應用面之統計分析，確認範例典籍中的Metadata。在實作方面，本研究以BNF來描述和定義Metadata，並以XML為工具完成雛型以供測試之。其中，本研究發現，基於資料倉儲觀點所擷取之後設資料的分析單位較傳統圖書典藏所得之為小。此外，本研究擷取過程中所涉及之Metadata，以功能性者為多，本研究亦採取了若干語言分析以期同時能維持典籍之文字結構。 / The objective of this research work is to acquisit and design Metadata for the construction of data warehouse of Traditional Chinese Medicine (TCM) literatures in the context of knowledge management. In order to solve the problem of preservation and utilization of TCM literatures, this work aims to designate the Metadata based on the viewpoint of knowledge engineering and data warehouse. In this work, the characteristics of the TCM regarding Metadata result in the 5W1H’s principle, while this work argues for its advantages for deriving more functional descriptions and keeping the syntax structure of the originals at the same time. To minimize the constraints of time, this work chooses “Ben Cao Bei Yao” and “Yi Fang Ji Jie” as the target to analyze. In constructing a prototype, the tacit knowledge in the example TCM literatures is converted through an analytic process explicitly into the organizational knowledge that can be easily preserved and processed by machines. Therefore, a statistical process is employed to derive and verify the Metadata in the context of the example TCM literatures. Then, the components regarding the Metadata are implemented with XML tools to develop the prototype. Last but not the least, this work presents its findings as follows: 1. The unit of analysis for deriving Metadata related to data warehouse is usually in a smaller degree of finesse in comparison to what is addressed in the area of traditional library management. 2. Through the Metadata derived in this work based on a data warehouse approach presents more functional elements, we can still maintain the linguistic structure of the example literatures with some careful linguistically analyses in the last step.

關係行銷或是一對一行銷是目前行銷領域上廣泛被討論的議 題，企業要如何透過有效的辨識、區隔、互動以及客制化來量身打造 顧客專屬的個人化產品與服務內容，並強化其重複消費動機及忠誠， 為目前各種產業爭相積極追求的目標，此外，由於微利時代風暴，各 產業無不希望透過顧客價值的辨識與經營，實現以更有效、更低的成 本的差異化行銷策略來創造高收益的企業經營目標，以航空產業如此 資本密集，高固定成本，低變動成本以及不對稱的供需平衡，誰掌握 低成本領導與差異化策略優勢，便能決戰存續於二十一世紀超競爭時 代之中。 由於資訊科技、網際網路以及資料探勘技術的臻於成熟， 充份 發揮了跨國、即時、深度滲透與互動的特性，使得關係行銷、一對一 行銷的實現變得更加有效而可行。本研究希望從顧客價值的認定、顧 客忠誠策略以及資料探勘技術的探討，來思考如何運用於航空公司會 員系統的顧客區隔，同時，希能透過航空公司產業通路架構、全球旅 行社訂位系統(CRS)的發展現狀、微妙的航空公司間策略聯盟以及不 同航空公司所提供的會員酬賓計劃內容的探討與陳述，初略地對個案 公司的所在環境進行策略性分析，以建議其所需採取投入關係行銷的 主要焦點客層。 緊接著， 利用資料探勘工具中的分群技術， 選定有效的指標變 數，針對某一區間的會員交易資料進行分群，藉由研究各群會員所蘊 含的特殊屬性，如營收貢獻、產品特性、通路喜好以及消費行為等等， 依據前述所定義的目標客層，以創造顧客價值為目標，精確建立目標 客戶群，並據以設計不同的行銷策略與產品組合，逐步深耕建立完整 會員關係行銷資料庫。 最後， 對於本研究所無法觸及的研究議題， 概略指出後續可能 的研究方向與建議。 / Customer Relationship Management and data mining in this hyper-competitive era have revealed a lot of interesting and innovative opportunities to enrich the capability of company to realize and provide customer value. They touch the most critical issue of the enterprise, “How can we create and sustain successful advantage, and maximize profitability by leveraging new technologies ?＂In this thesis, we will focus on the application of data mining in the FFP of the airlines industry, and look over the differences among FFP members to discover the implicative needs of FFP customers. First of all, we start discussion on literature review in chapter two, which was divided into three parts: customer loyalty strategy, customer value and data mining. In this chapter, we put emphasis on the concepts and definitions of above topics, and they would be helpful to us to select and decide key variables in the following data mining practice. Chapter three of this thesis is to introduce the structure and characteristics of the airlines industry, the history of Computerized Reservation System(CRS), the airlines strategy alliance and the FFP system, and to figure out the way to understand the existing threats and opportunities. Chapter four, which was abode by the steps of data mining process, defines business issues and collects around one year's FFP historical transaction data to establish the target data and perform an actual data mining practice. In this real practice, we use the demographic cluster function of IBM Intelligent Mining tool to do member clustering. We select net revenue, first and business class spending rate, reservation booking designator and customer activation as analytical variables to perform FFP member clustering. Each variable has been well equipped with weight and method to produce best cluster pattern. Finally, according to the mining results we have explored and interpreted, we provide our draft recommendations about marketing planning and mix activities from the perspectives of FFP members clustering.

參考Yue等論文，我們延續探討在通訊網路中對於非及時Available Bit Rate(ABR)資料與及時Variable Bit Rate(VBR)資料共同使用一條傳輸的等候模型。假設及時VBR資料傳輸較非及時ABR資料傳輸有優先權，我們建立與分析資料在離開過程之模型。本論文研究在非強制性優先權策略下之一般等候模型，藉由ABR資料的等候區域設立檢查點來控制流入量，在離開過程中推導VBR資料與ABR資料的離去時間的數學關係式，以及調查他們的數值模擬的表現。在此數學推導中需要藉MMBP關係，從ABR資料的等候區域觀點製造Markov矩陣算出穩定狀態下的機率分量、生成函數以及閒置時間的函數。結果發現檢查點影響兩者的離去時間並不顯著，但是VBR資料流入的速度卻會造成影響。 / Following the work by Yue et al., this thesis considers the departure of a multi-traffic network system for a popular communication network where a transmission link is shared by an Available Bit Rate (ABR) application for non-real time traffic and a Variable Bit Rate (VBR) application for real time traffic. It is assumed that the VBR traffic has a higher transmission priority than the ABR traffic. In this thesis, we establish a tractable analytical model of departure processes for such a system. The departure process is characterized by a general queueing model with a non-preemption policy for which the inter-departure times of VBR and ABR are derived, respectively. Since the VBR traffic is only affected when ABR is in service, the analysis is given to describe the departures of ABR, and VBR traffics. Numerical results are conducted to illustrate the system performance with input control of ABR traffic.