現在有許多對於眼動軌跡與人在認知方面的研究，包括理解狀態以及感興趣的程度；其中，閱讀文章時的眼動軌跡是最常被討論及研究的題材。而本研究的目的就是希望探討讀者在閱讀時的眼動軌跡，與其感興趣程度之間是否存在關係。 / 本研究的特色在於，我們不用一般分析眼動時關心每個AOI（area of interest）上的眼動資料，而是希望將眼動資料以序列的方式分析，並且運用資料探勘的方法，找出眼動序列中區分感興趣程度的眼動軌跡特徵的片段。 / 透過對於眼動軌跡的分析，我們希望研究的結果，在未來可以運用在資訊檢索的領域上，成為一種有效的「隱含式回饋（implicit feedback）」的方式，以改善現有資訊檢索效能。 / Much research has been performed on the relationship between eye movements and human cognition, including comprehension and interesting degree. The purpose of our research is to find out if there are relationships between eye movements of reading and interesting degree. / Instead of analyzing the eye movements on each area of interest, the characteristic of our research is to transform eye movements to sequence data, and to determine the eye movement patterns which discriminate whether user is interesting or not by using the method of data mining. / Through the analysis of the eye movements, our research result can be used as one way of implicit feedback of information retrieval to improve the effectiveness of the search engine.

區間型資料一般定義為由一個連續型變數的上限及下限所構成，本文中，我們特別引進了一種與時間相關的區間型資料，在Irpino (2006, Pattern Recognition Letters, 27, 504-513)，他提出每個觀測值皆是由某個時段的起始值及終點值之有方向性的區間所組成，譬如某一支股票在某一周的開盤價和收盤價。過去已經有許多方法運用在區間型資料，但尚未有方法來處理有方向性的區間型資料，然而Irpino 延伸主成分方法來處理有方向性的區間資料。Irpino提出的方法以幾何學的觀點來說，可視為定義在多維度空間上對有方向性線段(一般都稱作“spaghetti”)的分析，在本文中我們有更作進一步的延伸，不僅引入股票的開盤價及收盤價，且引入當周的最高價及最低價來探索Irpino所遺漏的資訊。此外，我們也嘗試用貝他分配來取代Irpino所使用的均勻分配來檢測是否有明顯的改善。延伸的方法需要計算大量複雜的式子，包含了平均數，變異數，共變異數等，最後利用相關係數矩陣進行主成分分析。然而最後的結論為若考慮資訊的價值，以加入最高值和最小值的延伸方法是較好的選擇。 / Interval data are generally defined by the upper and the lower value assumed by a unit for a continuous variable. In this study, we introduce a special type of interval description depending on time. The original idea (Irpino, 2006, Pattern Recognition Letters, 27, 504-513) is that each observation is characterized by an oriented interval of values with a starting and a closing value for each period of observation: for example, the beginning and the closing price of a stock in a week. Several factorial methods have been discovered in order to treat interval data, but not yet for oriented intervals. Irpino presented an extension of principle component analysis to time dependent interval data, or, in general, to oriented intervals. From a geometrical point of view, the proposed approach can be considered as an analysis of oriented segments (nicely called “spaghetti”) defined in a multidimensional space identified by periods. In this paper, we make further extension not only provide the opening and the closing value but also the highest and the lowest value in a week to find out more information that cannot simply obtained from the original idea. Besides, we also use beta distribution to see if there is any improvement corresponding to the original ones. After we make these extensions, many complicated computations should be calculated such as the mean, variance, covariance in order to obtain correlation matrix for PCA. With regard to the value of information, the extended idea with the highest prices and the lowest price is the best choice.

華語流行歌詞一直是語言、文學、音樂或是文化研究等相關科系赤手可熱的研究題目，內容包含作詞者、修辭分析、風格、用韻及語言表達等，然由於歌詞數量龐大，難以全部以人工分析。 近年來，資訊科技日新月異不斷地進步，隨著Big Data議題受到注目，Data Mining在近年來相當熱門，然而針對華語流行歌詞的巨量資料探勘與分析研究並不多。因此，本論文研究以程式來自動化分析歌詞的樣式與特性，包括詞彙頻率、詞彙相鄰關係分析、歌名分析、使用語系分析、舊曲新唱、歌詞風格自動分類、用韻及修辭等，而研究資料係透過網路擷取知名網站內容，包含魔鏡歌詞網 (Mojim.com)、臺北之音HitFM聯播網 (www.hitoradio.com) 及教育部重編國語辭典，透過分析規則及以Non-Trivial Repeating Pattern等方法，來完成分析及系統實作。 透過華語流行歌詞的大量分析，探勘及了解各種歌詞的風格與特性，將可了解各種歌詞、作詞者的風格與特色，進而應用在歌詞資料的管理與查詢。 此外，本研究將八萬多首歌詞的各種分析資料設置成網站，提供予學術研究使用，希冀此研究資料能使華語流行歌詞相關研究研究，進行更深入地探討。 / Chinese popular music lyrics has been a popular topic for researchers who major in languages and literature, music or culture. Related studies include of lyricists, rhetoric methods, styles, rhyme and language expression. However, all these studies were performed by manual analysis. It is difficult to analyze large amount of lyrics manually. With advances in computer technology, big data and data mining techniques have been widely used in different kinds of data. However, to the best of our knowledge, none have been done on pattern mining from big data of lyrics of Chinese popular music. Therefore, the objective of this thesis is to discover patterns from tremendous lyrics data based on data mining techniques. We use data downloaded from www.mojim.com, http://dict.revised.moe.edu.tw/cbdic/ and http://www.hitoradio.com (Hit FM). Data mining methods are employed to find lyrics’ patterns and features, including frequent words, word adjacency, analysis of hit songs' names, lyrics’ language studies, cover song research, automatic style prediction, rhyme and rhetoric patterns. With the analysis of tremendous lyrics and data, the developed approaches of this thesis will be helpful for discovering distinguishing styles of lyrics and lyricists.

以往顧客服務對企業而言只是被動的支援單位，企業無法得知每一位顧客的想法，顧客的聲音也難以傳達到企業內部，然而資訊與通訊技術的進步打破了這樣的障礙，促使資訊透明化，企業藉由資訊科技的輔助可以直接接觸到每一個個別的顧客，了解顧客的想法與需求來帶動企業的運作，成為企業經營的競爭優勢。 顧客關係管理（customer relationship management；CRM）為近年新興熱門話題，國內外軟體大廠紛紛投入CRM市場，然而此環境尚未成熟，各家提出之系統功能相當不一致，本文提出企業在建置顧客關係管理環境時應具備七大單元，包括產品端資訊蒐集機制、互動機制、事件處理機制、儲存分析機制、內部鏈結機制、策略對應機制、回饋執行機制等，以供企業參考。 CRM重視區別出每一個個別顧客的屬性再提供客製化服務，本文針對北部某醫院進行訪談，探討其規劃中之CRM及進行狀況，驗證本文所提出模式之可行性。 / Because of the power of information and telecommunications technologies, business can keep tracking of their customers to know what they really want and how they actually use the product. Analyzing the information return from customers and products, business can provide active and accurate service to the right customer through the right channel at the right time and rise the customer satisfaction. 　The purpose of this article is to show a complete CRM model which includes product information collecting mechanism, interact mechanism, event processing mechanism, storage and analysis mechanism, internal linkage mechanism, decision support mechanism, feedback implementing mechanism. When business constructs the CRM environment, they can apply this model to their organization. 　This article also studies a hospital case and plans a future framework for it by applying the model.

隨著資訊時代的來臨，網際網路用戶數的急速成長，客戶資料大量湧入的結果造成了企業回應客戶的困難；此外，入口網站的日漸普及與市場新利基的建立，使得上網者對專業化網站的需求提高；而配合新行銷時代的來臨，新科技行銷已從以往的大眾行銷走向一對一的個人化行銷。在這樣的背景與動機驅動下，近年來資料倉儲技術的興起，在這個講求競爭及速度的時代，提供了一套真正能創造企業優勢的方法，資料倉儲與客戶關係管理也逐漸成為企業發展與競爭不可或缺之工具。 　本研究乃是利用資料倉儲等相關技術，期望能達到網路市場區隔之目的，進而提供行銷相關之建議。因此，研究中定義了入口網站之會員特性分析模式，從中發掘出會員特性模式之分析結果，再配合入口網站功能階段分類表的服務內容，與研究中所訂之行銷模式，修改並制訂最後行銷策略上之參考建議。 　目前國內在資料挖掘上之相關研究，乃以金融、銀行、保險等領域的應用較多，對於入口網站此領域的應用尚無相關研究出現。另外，在網際網路發達，且入口網站百家爭鳴的背景之下，如何提高會員註冊率，與如何保留住已註冊的會員，是網站經營者所需考量的兩個重要課題。有鑑於此，本研究期望帶給網路業者一個新的思考方向與啟發，也試著將資料挖掘的技術帶入網際網路的應用領域，由網站會員特性的初步分析開始，做一簡單之示範。 / With the approach of informational era, the Internet users grow rapidly and the mass production of customers' data results in the difficulty of the response from industries to customers. Moreover, the widespread use of portal sites and the establishment of new market opportunity let the Internet users raise their demand to specialized websites. Because of the coming of new selling era, the selling of new science and technology has changed from the former mass selling to one-to one individual selling. Under the drive of this kind of background and motive, Data Mining technology rises and develops recently. In the era that emphasizes competition and speed, we provide a method that creates the superiority of industries. Data Warehouse and Customer Relationship Management (CRM) is becoming the essential tool for the development and competition of the industries. 　This research uses some technologies related to Data Warehouse in order to segment the Internet market, and provide the related selling suggestions. Therefore, the research defines the analytic model of characteristics of members of portal sites, discovers the analytic results, and then to cooperate with the services mentioned in the classified table of functions/stages of the portal site and the selling model made in the research, and then modifies and makes the final suggestions on selling tactics. At present, the internal researches about Data Mining are usually applied in the domains of finance, banking, and insurance, not in the domain of the portal sites. Besides, in the situation of prosperous Internet and competition of portal sites, how to raise the register rate of members and how to keep the registered members are the two important topics considered by website managers. Consequently, this research hopes to bring not only a new direction and inspiration to managers, but also a new applied domain, Internet, for Data Mining. This research will begin with a simple example that analyzes the characteristics of a Portal Sit's Members and the Making of Selling Tactics.

面對全球資訊網(WWW)應用蓬勃發展，網際網路上充斥著各種類型的資訊資源。而如何有效地管理及檢索這些資料，就成為當前資訊管理的重要課題之一。在發掘資訊時，最常用的便是搜尋引擎，透過比對查詢字串與索引表格(index table)，找出相關的網頁文件，並回傳結果。但因為網頁描述資訊的不足，導致其回覆大量不相關的查詢結果，浪費使用者許多時間。 為了解決上述問題，就資訊搜尋的角度而言，本研究提出以文字開採技術實際分析網頁內容，並將其轉換成維度資訊來描述，再以多維度資料庫方式儲存的架構。做為改進現行資訊檢索的參考架構。 就資訊描述的角度，本研提出採用RDF(Resource Description Framework)來描述網頁Metadata的做法。透過此通用的資料格式來描述網路資源，做為跨領域使用、表達資訊的標準，便於Web應用程式間的溝通。期有效改善現行網際網路資源描述之缺失，大幅提昇搜尋之品質。

本篇論文基於市場上所存在的一些交易機制，探討漲跌停前後之股價行為。因為證券市場上存在一些交易規則，例如漲跌停限制、買賣價差、最小升降單位限制、競價制度等，這些交易規則，具有法定的效力，理所當然地會影響投資人的行為。這種以各種交易機制的存在，探討價格形成的過程，就是市場微結構理論之研究範疇。 本篇引用Hausman, Lo, and MacKinlay (1992)所建立之Ordered Probit模型來分析漲跌停前後之股價行為，以個股逐筆交易的價格變動為因變數，而建立因變數為間斷型之分析模型，並以等待撮合時間、交易量、落後期交易價格、買賣價差等經濟變數，來探討個股逐筆交易價格變動的成因。在此同時，鑑於以往研究多假定價量關係為線性，本研究引入非線性的概念，檢定價量之間是否存有非線線性之關係；最後，為使模型更具解釋力，我們引入異質性變異數。 第一章 緒論……………………………………………………………..1 第一節 研究動機……………………………………………..1 第二節 研究目的……………………………………………..7 第三節 研究範圍與限制……………………………………..7 第四節 研究架構與內容……………………………………..8 第二章 文獻回顧……………………………………………………….10 第一節 非同時交易………………………………………….10 第二節 最小升降單位……………………………………….11 第三節 買賣價差…………………………………………….14 第四節 漲跌停限制………………………………………….15 第五節 重要模型回顧…………………………………….…18 2.5.1 Chou(1996)……………………………………..18 2.5.2 Hausman, Lo, and MacKinlay(1992)…………..20 第三章 實證模型設定………………………………………………….25 第一節 資料來源…………………………………………….25 第二節 樣本選取…………………………………………….25 第三節 模型設定…………………………………………….26 3.3.1 價格的變動區間……………………………….26 3.3.2 解釋變數……………………………………….29 3.3.3 條件變異數的型式…………………………….32 3.3.4 價格與成交量之間非線性關係的檢定……….32 第四節 資料處理…………………………………………….33 第四章 實證分析……………………………………………………….36 第一節 模型基本統計分析………………………………….36 第二節 價量非線性關係的檢定…………………………….39 第三節 Ordered Probit模型實證分析……………………….40 第五章 結論與建議……………………………………………………..48 第一節 結論…………………………………………………..48 第二節 建議…………………………………………………..49 參考文獻…………………………………………………………………..50 / This thesis is an application of the market microstructure theory’. In light of some trading mechanisms in our stock market, such as price limit, bid-asked spread, tick size, and auction system, those trading rules would influence the behavior of investors. We want to study the process and outcomes of stock price under those explicit trading rules. We use the Ordered Probit model (Hausman, Lo, and MacKinlay, 1992) to investigate the stock behaviors when it hits price limits. We also use price change as the discrete dependent variable, and time elapsed, trading volume, lag price changes, bid-asked spread as explanatory variables. In order to make the model more explainable, heterogeneity is applied. Moreover, we also want to find out if there is any nonlinear relationship between price change and trading volume.

XML (eXtensible Markup Language)是今日新興在網路上所使用的延伸性的標記語言。它具有豐富的語意表達及與展現方式獨立的資料獨立性。由於這些特性，使得XML成為新的資料交換標準並且其應用在資料庫中產生了許多新的研究議題在資料儲存和查詢處理上。在本篇研究中，將研究XML資料庫中的績效評估的議題並且發展一個可適用於不同應用領域及各種平台上的使用者導向且開放型工作量模式以評估XML資料庫績效。此XML開放型的工作量模型包含三個子模型─XML資料模型、查詢模型以及控制模型。XML資料模型將模式化XML文件中階層式的結構概念；查詢模型包含了一連串查詢模組以供測試XML資料庫的處理查詢能力以及一個開放型的查詢輸入介面以供使用者依照需求設定所需的測試查詢；控制模型中定義了一連串變數以供設定績效評估系統中的執行環境。我們發展此系統化且具開放型的工作量方法可以幫助各種不同應用領域的使用者預測及展現XML資料庫系統的績效。 / XML (eXtensible Markup Language) is the emerging data format for data processing on the Internet. XML provides a rich data semantics and data independence from the presentation. Thanks to these features, XML becomes a new data exchange standard and leads new storage and query processing issues on database research communities. In this paper, the performance evaluation issues on XML databases have been studied and a generic and requirement-driven XML workload model that is applicable to any application scenario or movable on various platforms is developed. There are three sub-models in this generic workload model, the XML data model, the query model, and the control model. The XML data model formulates the generic hierarchy structure of XML documents and supports a flexible document structure of the test database. The XML query model contains a flexible classical query module selector and an open query input to define the requirement-driven test query model to challenge the XML query processing ability of the XML database. The control model defines variables that are used to set up the implementation of a benchmark. This open, flexible, and systematic workload method permits users in various application domains to predicate or profile the performance of the XML database systems.

空間認知(Spatial Cognition)指出人所理解的空間複雜度，也就是人與環境互動的過程中，經由記憶與感官經驗，透過內化與重建產生物體在空間的關係認知。認知圖(Cognitive Map)是最常被使用在評估空間認知。分析學生所畫的認知圖有助於老師們瞭解學生的空間認知能力，進而擬定適當的地理教學設計。我們視空間認知發展的學習歷程檔案是由這些認知圖所構成。隨著數位學習科技的進步，我們可以透過探勘認知圖的方式，探討空間認知發展的學習歷程檔案。因此，我們藉由透過圖像的空間資料探勘，分析學生空間認知發展的學習歷程。 空間資料探勘(Spatial Data Mining)主要是從空間資料庫或圖像資料庫中找出有趣且有意義的樣式。在論文中，我們介紹一種空間樣式(Spatial Co-orientation Pattern)探勘以提供空間認知發展學習歷程的分析。Spatial Co-orientation Pattern是指圖像資料庫中，具有共同相對方向關係的物體(Object)常一起出現。例如，我們可以從圖像資料庫中發現物體P常出現在物體Q的左邊，我們利用二維字串(2D String)來表示物體分佈在圖像中的空間方向關係。我們透過Pattern-growth的方法探勘此種空間樣式，藉由實驗結果呈現Pattern-growth的方法與過去Apriori-based的方法[14]之優缺點。 我們延伸Spatial Co-orientation Pattern的概念至時空資料庫(Spatio-temporal Database)，提出從時空資料庫中，探勘Temporal Co-orientation Pattern。Temporal Co-orientation Pattern是指Spatial Co-orientation Pattern隨著時間的變化。論文中，我們提出兩種此類樣式，即是Coarse Temporal Co-orientation Pattern與Fine Temporal Co-orientation Pattern。針對此兩種樣式，我們提出三階段(three-stage)演算法，透過實驗分析演算法的效率。 / Spatial cognition means how human interpret spatial complexity. Cognitive maps are mostly used to test the spatial cognition. Analyzing cognitive maps drawn by students is helpful for teachers to understand students’ spatial cognitive ability and to draft geography teaching plans. Cognitive maps constitute the portfolios of spatial cognitive development. With the advance of e-learning technology, we can analyze portfolios of spatial cognitive development by spatial data mining of cognitive images. Therefore, we can analyze portfolios of spatial cognitive development by spatial data mining of images. Spatial data mining is an important task to discover interesting and meaningful patterns from spatial or image databases. In this thesis, we investigate the spatial co-orientation patterns for analyzing portfolios of spatial cognitive development. Spatial co-orientation patterns refer to objects that frequently occur with the same spatial orientation, e.g. left, right, below, etc., among images. For example, an object P is frequently left to an object Q among images. We utilize the data structure, 2D string, to represent the spatial orientation of objects. We propose the pattern-growth approach for mining co-orientation patterns. An experimental evaluation with synthetic datasets shows the advantages and disadvantages between pattern-growth approach and Apriori-based approach proposed by Huang [14]. Moreover, we extend the concept of spatial co-orientation pattern to that of temporal patterns. Temporal co-orientation patterns refer to the change of spatial co-orientation patterns over time. Two temporal patterns, the coarse temporal co-orientation patterns and fine temporal co-orientation patterns are introduced to be extracted from spatio-temporal databases. We propose the three-stage algorithms, CTPMiner and FTPMiner, for mining coarse and fine temporal co-orientation patterns, respectively. An experimental evaluation with synthetic datasets shows the performance of these algorithms.

近年來海內外企業舞弊的新聞層出不窮，面對公司不斷發生的舞弊案件，必須思考、研究如何改變傳統稽核方法與工具，消極目的在協助董監事評估內控制度是否持續有效，積極目的在於執行舞弊專案稽核，進而強化公司之控制環境，逐步達到事前分析舞弊行為徵兆之可能。 本研究擬就資料倉儲之導入，建構個案集團公司內部稽核專屬資料倉儲，及藉由資料採礦分析軟體運用分析研究，挖掘出有價值的稽核元件知識；進一步完善持續性稽核的運作環境；為本研究動機所在。本研究將以跨全球經營之個案電子業集團公司為研究對象，從資料倉儲及資料採礦導入工作執行過程及內部稽核作業實際程序及分析結果加以研究，以下列議題進行相關分析探討： 議題一、個案電子業集團公司資料倉儲導入過程探討及成果分析 議題二、個案電子業集團公司導入資料採礦於內部稽核作業之運用分析 本研究較傳統單一議題或產業研究更具廣度，進而對更多在多面向內控舞弊查核議題上有興趣更深入探討的研究者，提供一個參考意見與研究方向。而以個案集團公司導入資料倉儲及資料採礦對內控循環作業中所涉及之集團內常見舞弊高風險交易作業(如：圍標、拆單等)的成果分析，廣泛運用於集團眾多事業群處及作業系統資料庫，是為本研究較為特殊之貢獻。 本個案研究係採取實地稽核作業實施方式，驗證資料倉儲及資料採礦導入分析結果。並歸納出多面向內控循環作業之舞弊查核、形成以風險評估為基礎之稽核計畫、專案舞弊查核結果再強化資料倉儲及資料採礦元件技術等研究結論。 / There is a competitive circumstance for all companies to face how to survive in the world, it includes the opening the marketing, competing with others, preventing from the internal fraud …and so on. Here is a study for the board of directors to focus on how to execute the internal auditing program to make everything perfect under control. To review lots of the prevailing fraud cases in past, a company must have its own opinions to adjust the suitable auditing program and to adopt the accurate detecting processing in order to strengthen the controlling power and narrow down the risk in advance. This study is how to establish the data input and then retrieve the useful information for further analysis based on the mega data bank in the prevailing system of one of public electronic companies in Taiwan. A company engages in the research and development of products, they can, pursuant to their need, add the research and development circle, including the control operation for the basic research, the design of products, the research of technology, the making and testing of products, the research/ development information, and the recording and storage of information in system. There are two main factor issues for further discussion and break it down as the following--- (1) The investigation on the practical data input and the analysis by the result (2) How to retrieve the data for the internal auditing purposes This subject study is different from the same case of the past in the electronic industry in Taiwan. If you scrutinize the details of this report, you will find out lots of fraud cases in the company and the accurate auditing program taken for each of them. There are very interesting and ambitious directions for the public people to consider and investigate in order to fit it in their own industries in future. There are lots of frauds happened in the progressing of the global biding or local biding. You are able to detect the fraud by the internal control points through the data analysis from the prevailing computer system. The result in group biding or separating biding, those activities are existed under controlled by the certain people for benefit themselves. This is a true case happened in Taiwan. The internal auditor of this company would adopt the appropriate auditing progressing by verifying the data warehousing and data mining. A service enterprise’s internal audit unit shall, based on the results of the fraud detecting, the risk assessment, to renew the auditing plan stipulated and designed by the subjected factor into the book.