由於近年來大家對於「保險」一詞之觀念由負面逐漸轉為正面的保障,且因 為金管理機構的開放,使得保險業者成立家數變動幅度大,且由於法規鬆綁,能 夠提供之保險產品類型與銷售通路亦逐漸多元化,導致保險業者競爭激烈。而隨 著電腦技術進步及資料採礦技術蓬勃發展,許多公司積極投入資源,企圖利用資 料採礦技術從龐大資料中挖掘出新發現,藉以提供有用的資訊,作為公司決策的 依據,為公司創造出新的商機。因此善用現有資源,針對特定族群予以最適當及 最能滿足其需求的商品是保險業者最重要的目標。本研究期望針對最具潛力的族 群-年輕族群,利用資料採礦技術中之集群分析將其分群,並統整歸納出群集內 的共同特徵或特性,藉此描繪出不同類型的族群以更瞭解其需求。並利用關聯分 析法分析族群內保險商品購買情形,以做為保險業者針對此年輕客群中的不同族 群間保險商品之商品組合及未來商品規劃之策略建議。 本研究結果之總結發現,此年輕客群中上可分類成三個子集群,而三個之間 的特徵描述如下:A 集群:低風險、高忠誠度、重視退休養老生活的人。B 集群: 高風險、低忠誠度、重視身體健康的人。C 集群:高風險、低忠誠度、具投資理 財觀念的人。而根據此集群分類後之結果,利用關聯分析找出其保險商品最適合 之搭售組合。

在面對市場強力競爭之下，許多企業為了達到永續經營的目的，往往藉由改善產品品質、降低生產成本以及加強產品的彈性與效能，以便創造出符合顧客需求的優良產品，進而提升市場競爭力。 本研究主要的對象為電源轉換器(Switch Power Supply)。該電源轉換器在運轉的過程中時常會有溫度過高的情況發生，進而影響顧客對於產品的滿意程度，因此希望藉由降低電源轉換器的溫升以及溫升變異，來提升產品的品質以增加顧客的滿意度。在本研究中利用田口方法以及實驗設計去規劃出適當的實驗流程與實驗方法，並且經由實驗來收集實驗數據，分別採用灰關聯分析、主成分灰關聯分析、模糊評估分析和倒傳遞類神經網路等四種方法進行實驗分析，以決定出最適因子水準組合。 根據工程經驗與實驗結果得知，電源轉換器內主要發熱零件為IC、T1、LF1和D7。最適組合之確認實驗與現況比較發現，雖然LF1的平均溫升約比現況高2℃左右，但是IC、T1和D7的平均溫升卻可以降低2∼4℃，而且這四個主要發熱零件的溫升標準差也都有大幅降低的現象，由於降低產品變異也會提昇產品品質，一旦產品品質提升了便能夠增加市場競爭力，並且增加顧客的購買意願，因此本研究所找出的最適外殼鑽孔形狀與矽膠片厚度組合的改善效果良好。

電子商務，這是在90年代才興起的經營模式。在過去的社會演進裡，人類從最早的農業經濟進步為製造經濟，一切以產品、品質為出發；隨著知識時代的來臨，製造不再只是一味大量生產，更為重要的是站在顧客角度思考，而慢慢從製造經濟轉變為服務經濟，以顧客的滿意為重。 以滿足顧客的需求來看，網路商店必須具備高效率、可授權的、動態的且反應速度快的特性。消費者需要個人化資訊來做決定，然而，這似乎都是現有電子商務網站所欠缺的。 本研究將會採用資料探勘、灰關聯分析、層級分析法與灰色預測來達成客製化的行銷策略、產生客觀的產品排名與客製化的產品排名，並加以預測客戶的喜好。 / E-commerce, it’s a new business model from 90’s. On the social evolution track, from agriculture economy to manufacture economy, product and quality is the spotlight. With the coming of knowledge era, manufacturing is not just mass-productive, this era’s spotlight is the customer satisfaction. The evolution track had moved from manufacture economy to service-oriented economy. If sellers want to meet customers’ need, it should had some features, such as efficient, empower, dynamic, quick response, and so forth. Customers need tailor-made information to make the purchasing decision. However, nowadays internet stores cannot meet this need. This research will utilize data-mining, grey relation analysis, analysis hierarchy process and grey perdition to draw up tailor-made marketing strategies, generating objective product ranking and tailor-mage product ranking, and predict customers’ preference trend.

摘 要 隨著產業全球化競爭磅礡，傳統產業在面臨顧客導向的趨勢下，如何研發更具競爭力的產品與提升產品的品質為其重要課題。 本研究主要以北縣某電子股份有限公司之電源轉換器(Switch Power Supply)為研究對象。本產品在研發過程中時常發生零件溫度過高，導致無法符合顧客規格需求的情形。經由現況了解、探討影響電源轉換器溫度過高的關鍵因素和實驗數據的收集後，本文分別採用田口方法、灰關聯分析、主成份灰關聯分析、灰決策、多屬性損失函數、有規格界限的多屬性損失函數和倒傳遞類神經網路等七種方法進行分析，以決定出降低電源轉換器溫升之最適零件水準組合及再現性。 經由最適零件水準組合之確認結果得到各零件平均溫升降低約5~10℃且溫升標準差降低約3~6℃，使得各零件溫升符合國家安全規定，而且改善後總不良率與期望損失均降低近三成以上。各種結果皆證實本研究可使產品品質大幅提升，並相信未來在市場佔有率和顧客滿意度上也皆能顯著增加。 關鍵字：參數設計、田口方法、灰關聯分析、灰決策、主成份分析、多屬性損失函數。

近年來以電子資訊產業掛帥的台灣，傳統產業似乎成了被遺忘一個部分。但透過許多相關產業的調查資料顯示，部分傳統產業在全球化的競爭底下，依舊可在台灣立足，汽車業與船舶業就是其中的代表。由產業群聚的觀點來看，產業與其關聯產業在空間中的關係為何，產業是否集中於某些特定的空間單元，皆為值得討論的課題。 為對汽車業及船舶業進行分析，本研究透過產業關聯分析以及地理資訊系統之熱點分析（hot spot analysis），並結合工商普查資料、產業關聯表、生產者投入係數，進行空間集中指標的計算，從時間序列的變化，瞭解不同時間點運輸製造業的主要關聯產業之差異及其群聚版圖的變遷，並且進一步探討運輸製造業產業群聚之水平連結在空間臨接上所產生之差異。 研究結果顯示在1981年至2001年間，汽車業與船舶業在既有的空間單元中保持穩定的成長，僅北部區域的汽車業集中重心由台北移至桃園，此種情況即代表產業群聚的區位惰性。此外，船舶業在空間單元中有集中於鋼鐵業以及港口周邊的情況；汽車業與其主要關聯產業則都集中在桃園新竹一帶。 / In recent years, Taiwan taking electronics and information industries as main development, the traditional industry seems to become a part forgotten. But show through the survey materials of a lot of relevant industries, some traditional industries that under the competition of globalization, can still base on Taiwan, the automobile industry and shipping industry are representatives among them. In the view point of industry clusters, what are the spatial relationship between industry and its related industries, whether the industry concentrates on some specific space units, all in order to worth discussing. In order to analysis the automobile industry and shipping industry, this research passes the industry linkage analysis, hot spot analysis of geographic information system, and combine the industry, commerce and service census, input-output table, input coefficients table at producers' prices, to make the calculation of spatial concentration index, and from the view point of time series, to find out the difference of main related industries, the change of cluster map, and further more, to discuss the spatial relationship between industry and its main related industries. The result of study shows between 1981 and 2001, the automobile industry and shipping industry keeps steady growth with in the space unit that has already had, only the automobile industry of the northern area concentrates center on being transferred from Taipei to Taoyuan, this kind of situation represents the inert of location of industry clusters. In addition, shipping industry centre in nears the steel industry and port; the automobile industry and its main related industries are mostly concentrated in Taoyuan and Hsinchu.

隨著工業發展而產生的環境污染問題，促成了綠色供應鏈管理的提倡，尤其歐盟公佈危害物質禁用指令 (RoHS) 和廢電機廢電子設備指令 (WEEE) 的法令，讓企業對於環境保護的責任要盡更多心力。在現有供應鏈上加入綠色的考量，對於公司而言勢必是一種新的挑戰與壓力，這中間存在著許多的不確定性，風險也變得越來越複雜，在經營的過程中不可能完全去除風險，一個不利事件的發生會使得企業在營運、績效表現、成本、品質上，造成負面影響及損失，因此對於風險更應該站在事先預防的角度來管理它。如果能適當地評估，依據風險對於公司的重要性、潛在反應及其嚴重性，得知風險控管的優先順序，才不會將資源浪費在發生機率很小、影響程度不大的風險因子上。 綠色供應鏈的實施會關係到上下游廠商，是一種整合、全面性的作業活動，從上游的採購原料、製造過程、到下游的回收動作，都必須要有綠色的考量，需要彼此互相整合才能成功實施，這當中也會產生許多的問題。由於危害物質限用指令 (RoHS) 有規定八大類的電子電機產品限制使用鎘、鉛、汞、六價鉻等物質，會影響到採購和製造階段；而廢電子電機設備指令 (WEEE) 有規定回收率和與回收再利用率，會影響到企業的回收階段。因此從不同角度出發，找出企業在綠色採購、綠色製造、回收當中，不同階段中可能衍生出的風險因子，如此才能做到完整的考量。 本研究建立整體的綠色供應鏈風險評估模式，探討企業在實施綠色供應鏈可能所衍生的風險，找出影響的風險因子，並透過失效模式與效應分析(FMEA)，加入模糊語意概念、灰關聯分析的方式，針對風險因子可能所帶來的嚴重性、發生度、難檢度進行分析評估，藉由計算出的風險值，可以排列出風險因子的風險大小，讓企業能知道控管的優先順序。 / Along with the industrial development has brought the environmental pollution question. Therefore green supply chain management was promoted. European Union announces the WEEE directive (Waste Electrical & Electronic Equipment) and RoHS directive (The Restriction of the use of Certain Hazardous Substances in Electrical and Electronic Equipment). These regulations make manufactures & retailers take charge of environment protection. The green consideration will be subsumed into the supply chain. It’s a new challenge and pressure for company. Because green supply chain management implies uncertainty, the risk will be become more and more complex. The company couldn’t remove risk. The occurrence of harmful event may cause cost, quality, performance negative affection and financial failure. The company should manage risk in advance. If the company can appropriately assess risk, according to the importance, severity, occurrence of risk, the company will get the priority of risk control, we won’t waste our resources on the risk that the occurrence probability is very small and influences degree is not very great. To implement the green supply chain will influence supplier, manufacturer, recycler and Name Brand Vendor. It’s a conformability and comprehensive activity. From supplier purchase raw material to Name Brand Vendor recycle the product, they all need a green consideration. They need the mutual conformity, so they can successfully implement the green supply chain. The RoHS has already ruled the restriction that Electrical and Electronic Equipment are limited to use Lead, Cadmium, Mercury, Hexavalent Chromium, Polybrominated Biphenyls, etc, and therefore will influence the purchase stage of business and the manufacturing stage of business. The WEEE regulation include recovery rate and recue/recycling rate, so will influence the return stage of business. We should assess risk from the different viewpoint, and find out risk factor of the purchase stage, the manufacturing stage of business, and therefore we can have a complete consideration. This research constructs the risk assessment model for green supply chain, and probes into influential risk factors when the company implements the green supply chain. This study based on Failure Mode and Effects Analysis (FMEA) that combine fuzzy set theory and grey theory, to analyze severity, occurrence, detection of risk factor. Results of this study show that risk value of risk factor, the arrangement of risk factors and the priority of risk control.

全球面臨人口老化與人口不斷成長的壓力下，對於醫療服務的需求不斷提升。醫療服務領域中常以資料探勘「關聯規則」分析，挖掘隱藏在龐大的醫學資料庫中的知識(knowledge)，以支援臨床決策或創新醫療服務。隨著醫療服務與應用推陳出新(如，電子健康紀錄或行動醫療等)，與醫療機構因應政府政策需長期保存大量病患資料，讓醫療領域面臨如何有效的處理巨量資料。 然而傳統的關聯規則演算法，其效能上受到相當大的限制。因此，許多研究提出將關聯規則演算法，在分散式環境中，以Hadoop MapReduce框架實現平行化處理巨量資料運算。其相較於單節點 (single-node) 的運算速度確實有大幅提升。但實際上，MapReduce並不適用於需要密集迭帶運算的關聯規則演算法。 本研究藉由Spark記憶體內運算框架，在分散式叢集上實現平行化挖掘多維度多顆粒度挖掘關聯規則，實驗結果可以歸納出下列三點。第一點，當資料規模小時，由於平行化將資料流程分為Map與Reduce處理，因此在小規模資料處理上沒有太大的效益。第二點，當資料規模大時，平行化策略模式與單機版有明顯大幅度差異，整體運行時間相差100倍之多；然而當項目個數大於1萬個時，單機版因記憶體不足而無法運行，但平行化策略依舊可以運行。第三點，整體而言Spark雖然在小規模處理上略慢於單機版的速度，但其運行時間仍小於Hadoop的4倍。大規模處理速度上Spark依舊優於Hadoop版本。因此，在處理大規模資料時，就運算效能與擴充彈性而言，Spark都為最佳化解決方案。 / Under the population aging and population growth and rising demand for Healthcare. Healthcare is facing a big issue how to effectively deal with huge amounts of data. Cased by new healthcare services or applications (such as electronic health records or health care, etc), and also medical institutions in accordance with government policy for long-term preservation of a large number of patient data. But the traditional algorithms for mining association rules, subject to considerable restrictions on their effectiveness. Therefore, many studies suggest that the association rules algorithm in a distributed computing, such as Hadoop MapReduce framework implements parallel to process huge amounts of data operations. But in fact, MapReduce does not apply to require intensive iterative computation algorithm of association rules. Studied in this Spark in-memory computing framework, implemented on a distributed cluster parallel mining association rules mining multidimensional granularity, the experimental results can be summed up in the following three points. 1th, when data is small, due to the parallel data flow consists of Map and Reduce, so not much in the small-scale processing of benefits. 2nd, when the data size is large, parallel strategy models and stand-alone obviously significant differences overall running time is 100 times as much when the item number is greater than 10,000, however, stand-alone version cannot run due to insufficient memory, but parallel strategies can still run. 3rd, overall Spark though somewhat slower than the single version in small scale processing speed, but the running time is less than 4 times times the Hadoop. Massive processing speed Spark is still superior to the Hadoop version. Therefore, when working with large data, operational efficiency and expansion elasticity, Spark for optimum solutions.

不同股市的報酬關聯隨時間動態改變，本研究欲了解近期美國、台灣與亞太地區的中國大陸、香港、日本及韓國的報酬連動關係，並進一步觀察次級房貸風暴期間美股對這些地區的關聯改變趨勢。本論文採用灰色理論與時間序列兩種方法，實證發現次級房貸風暴發生期間，台股及亞太地區主要指數不論在報酬率或是報酬率波動性受美股影響的程度大多增強。 實證結果顯示，在風暴期間的報酬率傳導關係，亞太以韓國影響台股最顯著，美股則全面影響亞太指數；在報酬率波動性溢傳上，亞太以日本、美股以道瓊工業影響台股最強，台股則是電子類股被美股影響最重，但營建類股在與美股或是亞太指數的關聯趨勢變化卻最明顯。另外，灰關聯分析對時間序列檢定的關聯組合可以提供互補的關聯強弱關係說明，且具有相當的正確性。 / Connections between stock markets are dynamically changing, and it affects investor's transnational investment portfolio. We focus on the relationships of stock markets among the United States, Taiwan, Japan, Korea, China and Hong Kong, and eager to understand the connection tendency between Untied States and Asian-Pacific area during the subprime mortgage crisis period. The identified research methods are time series and grey theory, including Granger causality test, GARCH model and grey relational analysis. We find out the returns and volatility in Asian-Pacific stock markets were all affected increasly by U.S. market during the subprime mortgage crisis. The main empirical results are as follows: In the relationships of returns, Korea affects Taiwan mostly in the Asian-Pacific area, and U.S. market affects all the others entirely during the subprime crisis. In the relationships of volatility, Japan and Dow Jones index affects Taiwan deeply during the period; within all the Taiwan indexs, Electronic Sector Index was affected by the U.S. market mostly than others during the same period, but the connection tendency in the Construction Sector Index with other markets changes more obviously. Otherwise, grey relational analysis can provide complementary explainations as compared to time sereies in the strength of relationships, and the explainations are with plenty credibility.

房地產業的活動被一般人認為是「火車頭產業」，探究這種未經學術嚴謹定義的說法，涵意概有兩層：其一是認為房地產業有極大的「向後關聯」效果，將可帶動相關總體經濟產業的發展。另一層涵意是指房地產業既然有帶動總體經濟繁榮成長的功能，也就意味著房地產業活動所構成的房地產景氣具有領先總體經濟景氣的特質，而為一般景氣昇沉的預期訊號。惟這種說法似乎與現實情況不合：現總體經濟景氣已有復甦跡象，但房地產業卻相對的毫無起色，因此本研究從「房地產業對總體經濟活動之影響分析」、「房地產景氣與總體經濟景氣在時間上領先、同時、落後關係之探討」兩部份，分別以較嚴謹的「產業關聯分析法」與「景氣綜合指標分析法」，來探討該說法的正確性及政策等含意，獲得「尚無充份的證據支持房地產業是火車頭產業」的結論。 　　有關政策涵義方面：房地產業的向後關聯效果不強，意味著政府如意圖以房地產業為振興經濟的逆循環政策應改變至回歸市場機制的調控，而不應有太多的政策介入。政府不必因總體經濟的不景氣而企圖刺激房地產景氣；亦無須強調總體景氣過熱而打壓房地產景氣。至於「房地產景氣與總體經濟景氣在時間上領先、同時、落後關係」部份，不論房地產綜合、各層面、基準循環指標之景氣與總體經濟綜合、構成房地產綜合景氣重要指標時間上關係比較中，我們有足夠的證據認為「房地產景氣落後總體經濟景氣」。在預測上的涵意是若干重要總體經濟指標可以用來預測房地產景氣未來的走勢。 / The fluctuation in the real estate market is of long-standing, and has evoked much discussion, particularly how the real estate activities and cycles are related to macroeconomics has been an important issue drawing tremendous attention in Taiwan. This research contains two parts : in the first part, we have applied the method of lnput-Output(I/O) analysis to identic the backward linkage of the real estate sector. In the second part, we try to use the method of composite indexes of business cycle for real estate cycle indicators, including individual activities, four different stages of real estate life cycle -- investment, construction, transaction, and utilization, to clarify the " timing " relationship between business cycle and real estate cycle. 　　Based on the economic analysis, the results of this research are following : 　　1. We have not found strong evidence supporting the important backward linkage of the real estate sector. It means, in the view of using real estate activities for pushing macroeconomics, the government should not intervene the activities of real estate industry to market mechanism due to the effect of real estate activities contribute little feedback to macroeconomics. 　　2. Our investigation reveals the macro-variables, such as GDP, M2, the index of stock market, CPI, composite index etc. , tend to be leading indicators of real estate activities over twelve months approximately. This means, in the view of forecasting, we can use certain macro-variables to forecast the trend of real estate cycle in the fliture.

本篇論文旨在研究中文流行音樂歌詞與歌曲之間情意的關聯性，並設計一個能推薦出符合歌曲情意的「以曲找詞歌詞推薦系統」。 流行音樂（Popular Music）在廣義上的定義為透過大眾媒體傳播、以大眾為閱聽對象的歌曲。其大眾化的特徵，使得流行音樂歌詞的主題多與日常生活息息相關且能清楚表達歌曲的情意，並以其所引起的共鳴性決定歌曲是否具出版的商業價值，人們也常常使用流行音樂歌曲來唱出屬於自己的故事、屬於自己的心聲。因此，本篇論文提出自動為流行音樂歌曲推薦符合歌曲情意的歌詞，讓舊有的歌曲搭配上新的歌詞，而當一首歌曲搭配了不同的歌詞就有了不同的故事，也帶給了原曲新的生命，達成一曲多詞的數位加值效果。 由文獻及專業音樂創作者的論述中，我們可以了解流行音樂詞曲有相關的搭配關係，其中又以詞曲情意的搭配關係最為重要，因此詞曲情意之間的關聯性為本研究問題的核心所在。透過大量分析市面上的流行歌曲，我們便可以從中看出詞曲之間情意搭配的線索。我們利用 LSA（Latent Semantic Analysis）演算法萃取出歌詞的情意特徵值，並比較其與語言學領域中隱喻融合理論的相似性，而在歌曲方面萃取出音高、調性、速度、節奏、和弦及音色等與等能展現歌曲情意的相關特徵值。然後利用了 CFA（Cross-Modal Factor Analysis）演算法來建立詞曲之間情意特徵值的關聯模型，最後我們便可以利用關聯模型來建立推薦系統，如此便完成了詞曲情意關聯為基礎的以曲找詞歌詞推薦系統。 實驗結果顯示，考慮詞曲情意特徵關聯所訓練出的關聯模型（CFA Feature Model）在以曲找詞推薦符合情意歌詞的前五名準確率平均達 60.1 %，前五十名也有 41.4 % 的準確率，比起僅考慮歌曲情意特徵（Audio Feature Model）以曲找詞推薦符合情意歌詞的前五名準確率 45.1% 及前五十名準確率28.6 % 準確率高，代表本研究所提出的詞曲情意關聯模型確實能有效推薦出符合歌曲情意的歌詞。我們也對本研究提出的詞曲情意關聯模型進行歌詞推薦結果的案例分析，我們輸入幾首學生創作的歌曲觀察詞曲情意關聯模型歌詞推薦結果，我們發現推薦出的流行音樂歌詞與學生創作的原詞在歌詞情意上非常類似，再次顯示本研究所提出的詞曲情意關聯模型確實能有效推薦出符合歌曲情意的歌詞，在詞曲創作上將能為創作者帶來靈感支援，幫助創作者詞曲創作。 / Nowadays lots of people use popular music to sing out their own story, and their own aspirations. In this thesis, we propose an approach to analyze the conception association between lyrics and music of Chinese popular music. And for applications, we design a lyrics recommendation system which can automatically recommend lyrics which is suitable to accompany with query music according to the affection and conception between lyrics and music. So, the old song with new lyrics, just like the song with different stories, brings the original song with new life. There are accompany association between lyrics and music, and the affection and conception association is most important among all. Therefore, analyze the conception association between lyrics and music is our goal. To do this, we can find out the association clues between lyrics and music from analyzing lots of popular music. For lyrics, we use LSA (Latent Semantic Analysis) algorithm to extract lyrics conception features. For music, we extracted the pitch, tonality, speed, rhythm, chords features which can show the music’s conception in the music. Then we use the CFA (Cross-Modal Factor Analysis) algorithm to analyze and learn the conception association between lyrics and music and establish the conception association model . Finally, we will be able to take advantage of the conception association model to establish the lyrics recommendation system. In the experimental results, when recommend the same conception lyrics to the query music, our proposed approach (CFA Feature Model) reaches accuracy of 60.1% on average in the top 5 recommended lyrics. Compared to control group approach (Audio Feature Model) only reaches accuracy of 45.1% on average in the top 5 recommended lyrics, our approach get better accuracy. We also presented some interesting lyrics recommendation results in case study. We upload some popular music created by students, and we found out that the affection and conception of the recommended lyrics are similar to the original song lyric which is created by the students. The experimental results show that the lyrics and music conception association model we proposed in this study does recommended lyrics suitable to the query music conception.