傳統探討分析師推薦資訊價值的研究多採用累積超額報酬的方式，近年來研究顯示個別投資人的績效顯著低於機構投資人，因此是否分析師推薦能夠幫助提升個別投資人的福利。本論文的第一個貢獻在檢定是否個別投資人能夠獲取分析師推薦的資訊價值，為區分推薦資訊分別對於個別與機構投資人的價值為何，本研究採用的每種投資人實際的交易利潤作為衡量指標。 研究結果顯示所有投資人都可以透過買入推薦獲取顯著的正報酬，但在賣出推薦上，僅外資與共同基金仍能維持獲取正的報酬。同時發現在推 薦事件期間，專業機構投資人的利潤顯著高於一般散戶的獲利。 進一步，本論文的第二的主題在探討此推薦的資訊價值對於不同投資人的差異，是否肇因於推薦券商所提供的私有資訊，因此進一步將各類投資人分成推薦券商的客戶與非客戶。結果顯示國內機構投資人的利潤在客戶的身上顯著高於非客戶的獲利，顯示推薦券商在對外公佈推薦資訊前的確提供私有資訊給其國內機構客戶，但此現象在賣出推薦並不存在。 第三，本論文進一步分析是否拿到推薦券商所提供私有資訊的客戶也是推薦券商的經紀業務收益的主要貢獻者。在比較推薦券商與非推薦券商在被推薦股票上的相對交易量(金額)中，發現推薦券商的確因為買入推薦股票而增加經紀業務量，但很驚訝的發現貢獻最多交易量的是個別投資人，而非拿到最多好處的機構投資人。 最後，本研究透過迴歸分析探討不同投資人的交易利潤與推薦券商所獲得的經紀業務量的關係。在控制推薦類型、推薦評等與被推薦股票之股票特性後，發現投資人的交易利潤與推薦券商的經紀業務收益成正相關，再次顯示券商推薦與其各項業務收益間的關係。 / Traditionally, information value of analysts’ recommendations has been well-recognized by cumulative abnormal returns. Recent studies show that individuals are underperformed, and therefore, it is a critical issue on if analysts’ recommendations are helpful to individuals’ welfares. The first contribution of this dissertation to the literature is to examine whether individual investors are capable of capturing the information value. To classify the information value of recommendations for individuals and institutions, respectively, I, thus, use a direct measure to calculate the actual trading profits of types of traders. To our best knowledge, this is the first paper that demonstrates the information value for types of investors. Our results indicate that, all investors get positive and significant profits in brokerages’ buy recommendations, no matter what types of investors are measured. As to sell recommendations, only foreign investors and mutual funds have positive returns. We also find that professional institutions earn more profits than retail investors during the recommendation event periods. Further, the second objective of this dissertation is to test whether the information values are caused by private information from brokerages’ houses, we separate the profits of types of investors into customers and non-customers based. The findings are that only domestic institutional customers of recommending brokerages are more beneficial than those of non-recommending brokerages in buy recommendations, which implies that brokerage houses may reveal private information to their own institutional customers before buy recommendations make public. This does not hold for sell recommendations. Third, we are interested in analyzing whether the private information that recommending brokerages provide to their own customers may, indeed, contribute to brokerages’ commission revenues. By comparing the trading volume of recommending brokerages and non-recommending brokerage for the covered stocks, we find that the volumes of covered stocks issued in the recommending brokerages are increased for buy recommendations. Particularly, we find that the main contribution of trading volume is from individuals. Furthermore, we run regressions to study the relationship between trading profits of types of investors and trading volume of recommending brokerages. After controlling recommendation types, consensus rating of recommendations, and stock characteristics, our results indicate that trading profits of all types of investors are positively related to commission revenues of brokerages. This may justify the importance of brokerage recommendations on their business relationships.

隨著網誌成為網路世界的新寵兒，一種新型態的廣告「代言人網誌日記形式廣告」也應運而生。本研究以「代言人類型」與「置入方式」兩變項對此新類型的廣告之溝通效果三面向「回憶效果」、「品牌態度」與「購買意願」做初步的探討，並設計了三種不同的代言人類型與兩種置入方式共六種實驗廣告，於網路上徵求受測者並進行問卷之收集。 本研究結果發現如下：在「回憶效果」與「購買意願」兩應變項上，「不同類型代言人」與「不同型式置入」均會對其產生顯著影響，而在「品牌態度」應變項上，僅「不同類型代言人」對其產生顯著影響。同時，在此三面向之溝通效果上，均呈現出「一般消費者代言人」效果最佳、「名人代言人」效果最差的情形；在「回憶效果」與「購買意願」上，則均呈現出「文章中置入」形式效果優於「文章外置入」形式的情形。「代言人類型」與「置入方式」間的交互作用在此三面向之溝通效果上均未達顯著性。 / “Blog diaries written by product endorsers” is a new form of advertising promotion as blog becomes more popular in Internet world than ever. ”Endorsers types” and “product placement types” are the two independent variables in the study to investigate the three aspects of advertising communication effects—“recall effects”, “brand attitude”, and “purchase intention.”— which are the dependent variables in the study. The formal experimental commercials are six different kinds in total based on three endorser types and two product placement types and posted in Internet for questionnaires collecting. The results are both the different endorser types and the different product placement types have significant influences on “recall effects” and “purchase intention” aspects, but only different endorser types caused significant influences on “brand attitude” aspect. The typical consumer endorser is most effective type in all of three kinds and the celebrity endorser is the worst one on every advertising communication effect aspects. “The inside placement” has the better effect than “the outside placement” on “recall effects” and “purchase intention” aspects .The interact between “the endorsers types ”and “product placement types” are not significant in the three aspects of advertising communication effects.

在這個充滿繁多新媒體時代，使用者面臨到眾多資料和快速變動的環境，使用者在媒體的使用行為和選擇上更加依賴各種推薦平台的建議。除此之外，隨著社群媒體的興起，許多的推薦平台整合了社群的人們關係來提供更準確的建議和選擇。雖然推薦系統在影響使用者的使用行為有顯著的效果，然而企業和品牌卻鮮少去關注或了解如何增加顧客參與行為在整合社群媒體的推薦平台上。顧客參與行為並不只有傳統的交易行為，而是包含了所有直接和間接影響企業品牌的行為，像是使用者回饋、口碑傳播等。而且，現今尚未有清楚明確的定義哪些關鍵因素，會影響顧客參與行為在社群化推薦推薦系統，來藉此獲得顧客關注，形成正向生態系統。 本研究中，我們根據達人在社群化推薦平台中具有重要的影響力的觀點，以促進重塑達人知識來改變原有達人的行為和態度，藉此影響所有一般使用者在社群化推薦平台的顧客參與行為。我們提出新的架構和系統來幫助中小型商家在推薦平台上影響更多的推薦達人，獲得更多的顧客參與。我們建立商家參與後台來幫助中小型商家可以洞悉達人的行為，我們也建立了重新塑造資訊的系統，提供達人所需要的訊息文章，藉此來改變達人的知識和行為。此研究發現，達人的行為會受到娛樂型、知識型和激勵型的文章訊息影響行為，一般使用者也會受到達人行為影響。此外我們藉由品牌合作角度來幫助得到更多的顧客參與行為，我們發現中小型商家可以在社群化推薦平台獲得顧客參與且建立一個正向機制循環。 / With the highly dynamic trend of service economy, the firms are increasingly to co-create value with brand alliance to advance their competition advantage. On the other hand, with the massive information on the new media, the referrals provided by recommender systems in combination with social media have significantly impact on customer behavior. In light of these trends, the markers and firms should aim to increase the customer engagement behavior (CEB) which goes beyond the traditional transactions including purchase and non-purchase behavior on social recommenders. In this research, we focus on the role of mavens who are powerful influencers on the social recommender. We propose a new conceptual framework for facilitating to impact the maven’s knowledge and behavior and increase the CEB on the social recommender for Small/Middle Enterprise (SME). We establish the SME support engagement site for increasing the CEB on social recommender and framing knowledge context to influence maven for achieving the insight of the maven’s behavior. As the result of research, we discover that maven engagement behavior would be influenced by the entertainment, information and incentive types in context from the brand alliance perspective and the non-maven are willing to be affected by maven behavior. Moreover, with this discovery, the SME can increase the customer engagement behavior on the social recommender

現今許多網路商店或電子商務將產品銷售給消費者的過程中，皆使用推薦系統的幫助來提高銷售量。如亞馬遜公司(Amazon)、Netflix，深入了解顧客的使用習慣，建構專屬的推薦系統並進行個性化的推薦商品給每一位顧客。 推薦系統應用的技術分為協同過濾和內容過濾兩大類，本研究旨在探討協同過濾推薦系統中潛在因子模型方法，利用矩陣分解法找出評分矩陣。在Koren等人(2009)中，將矩陣分解法的演算法大致分為兩種，隨機梯度下降法(Stochastic gradient descent)與交替最小平方法(Alternating least squares)。本研究主要研究目的有三項，一為比較交替最小平方法與隨機梯度下降法的預測能力，二為兩種矩陣分解演算法在加入偏誤項後的表現，三為先完成交替最小平方法與隨機梯度下降法，以其預測值對原始資料之遺失值進行資料插補，再利用奇異值分解法對完整資料做矩陣分解，觀察其前後方法的差異。 研究結果顯示，隨機梯度下降法所需的運算時間比交替最小平方法所需的運算時間少。另外，完成兩種矩陣分解演算法後，將預測值插補遺失值，進行奇異值分解的結果也顯示預測能力有提升。 / Recommender system has been largely used by Internet companies such Amazon and Netflix to make recommendations for Internet users. Techniques for recommender systems can be divided into content filtering approach and collaborative filtering approach. Matrix factorization is a popular method for collaborative filtering approach. It minimizes the object function through stochastic gradient descent and alternating least squares. This thesis has three goals. First, we compare the alternating least squares method and stochastic gradient descent method. Secondly, we compare the performance of matrix factorization method with and without the bias term. Thirdly, we combine singular value decomposition and matrix factorization. As expected, we found the stochastic gradient descent takes less time than the alternating least squares method, and the the matrix factorization method with bias term gives more accurate prediction. We also found that combining singular value decomposition with matrix factorization can improve the predictive accuracy.

推薦系統為網站提升使用者滿意度、減少使用者所花費的時間並且替網站提供方提升銷售，是現在網站中不可或缺的要素，而推薦系統的研究集中在娛樂項目，學術研究論文推薦系統的研究有限。若能給予有價值的相關文獻，提供協助，無疑是加速進步的速度。 在過去的研究中，為了達到個人化目的所使用的方法，都有不可避免或未解決的缺點，2002年美國研究圖書館協會提出布達佩斯開放獲取計劃(Budapest Open Access Initiative)，不要求使用者註冊帳號與支付款項就能取得研究論文全文，這樣的做法使期刊走向開放的風氣開始盛行，時至今日，開放獲取對學術期刊網站帶來重大的影響。在這樣的時空背景之下，本研究提出一個適用於學術論文之推薦機制，以FP-Growth演算法與協同過濾做為推薦方法的基礎，消弭過去研究之缺點，並具個人化推薦的優點，經實驗驗證後，證實本研究所提出的推薦架構具有良好的成效。 / Recommendation system is used in many field like movie, music, electric commerce and library. It’s not only save customers’ time but also raise organizations’ efficient. Recommended system is an essential element in a website. Some methods have been developed for recommended system, but they are primarily focused on content or collaboration-based mechanisms. For academic research, it is very important that relevant literature can be provided to researchers when they conduct literature review. Previous research indicates that there are inevitable or unsolved shortcomings in existing methods such as cold starts. Association of Research Libraries purpose “Budapest Open Access Initiative” that is advocate open access concept. Open access means that users can get full paper without register and pay fee. It’s a major impact to academic journal website. In this space-time background, we propose a hybrid recommendation mechanism that takes into consideration the nature of recommendation academic papers to mitigate the shortcomings of existing methods.

「人」是公司中最重要的資產，而在知識密集的行業中，這樣的資產更顯得重要。由於網路技術的出現，網路人力銀行也成為另外一種人才招募的新興管道，但透過網路人力銀行所召募的人才素質並沒有傳統上透過公司員工推薦進來的人才可更進一步瞭解的好處。因此本研究透過一網路人才推薦信任制度，來加強線上人力銀行之人才篩選能力，希望透過此制度能繼續保有網路人力銀行在人才招募速度上的優勢，並能加強其篩選的能力。 本研究針對人才招募管道進行了文獻的探討，提出一人才推薦制度，以某國立大學EMBA之人才庫，透過成員間的學經歷背景相似度，推薦出擁有相同顯性工作能力的人才。讓人才招募單位可以得到推薦的人才，並可對其作信任評價的推論。接著利用實驗來求出雛型系統的一些關鍵參數，讓雛型系統運作得更完善以及更符合使用者的需求。 本雛型系統結合了網路人力銀行人才招募方式可快速地招募到大量員工的特點，及員工推薦人才招募方式可招募到更適切員工的特點。並透過FOAF格式的使用，將線上社會網絡的資料格式統一，有助於縮短整個人才信任推薦系統的建立時間。 / "Human Resource" is one of the most important assets of company, especially in knowledge-intensive industries. As network technologies developed, commercial job site has also become another kind of recruitment channel. But through this kind of channel, companies don’t have better chance to know new employee than traditional way. Therefore this study filters new employees by a Recommendation & Trust Inference mechanism. Hope that commercial job site would continue to keep the advantages of high efficiency in recruitment, and enhance its filtering capability at the same time. First, this study surveys literatures in recruitment channels. And it proposes a Recommendation & Trust Inference mechanism using a national university EMBA program member data as an example. The Recommendation mechanism recommend candidates having the same specialty by comparing their similarity of education and work experience. Furthermore, recruitment unit could use Trust Inference mechanism to get suitable candidates. Third, we conduct experiments to find the key parameters for the prototype system. Make the system able to work better and meet users’ needs. The prototype system combines the benefit of commercial job site which can quickly recruit a large number of employees and the feature providing more appropriate candidates for the company recommended by staff. Simultaneously by taking use of the FOAF format, we can unify the data format in online social network. The way mentioned above will effectively reduce the system set-up time.

近幾年在金融風暴及全球競爭等影響下，企業紛紛導入商業智慧平台，提供管理階層可簡易且快速的分析各種可量化管理的關鍵指標。但在後續的推廣上，經常會因商業智慧系統提供的資訊過於豐富，造成使用者在學習階段無法有效的取得所需資訊，導致商業智慧無法發揮預期效果。本論文以使用者在商業智慧平台上的操作相似度進行分析，建立相對於實體部門的凝聚子群，且用中心性計算各節點的關聯加權，整合至所設計的推薦機制，用以提升商業智慧平台成功導入的機率。經模擬實驗的證實，在推薦機制中考慮此因素會較原始的推薦機制擁有更高的精確度。 / In recent years, enterprises are facing financial turmoil, global competition, and shortened business cycle. Under these influences, enterprises usually implement the Business Intelligence platform to help managers get the key indicators of business management quickly and easily. In the promotion stage of such Business Intelligence platforms, users usually give up using the system due to huge amount of information provided by the BI platform. They cannot intuitively obtain the required information in the early stage when they use the system. In this study, we analyze the similarity of users’ operations on the BI platform and try to establish cohesive subgroups in the corresponding organization. In addition, we also integrate the associated weighting factor calculated from the centrality measures into the recommendation mechanism to increase the probability of successful uses of BI platform. From our simulation experiments, we find that the recommendation accuracies are higher when we add the clustering result and the associated weighting factor into the recommendation mechanism.

潛水是網路社群中的普遍行為，並且潛水者常為網路社群中的多數，通常潛水者從社群中獲取得多，但卻貢獻得少，雖然對於整體社群無害，但對於網路社群的貢獻卻相當有限，無助於整體社群的發展與成長。因此，如何激發潛水者更積極參與互動討論，樂於貢獻一己之力，對於網路社群的發展甚為關鍵。特別是在數位學習環境中，更應該積極發展有效激發潛水者策略，以促進潛水者更積極參與社群互動討論的意願，提昇整體社群合作學習動力。而透過讓潛水者感受到自己參與社群互動的重要，提昇潛水者的社會知覺，是否有助於激發潛水者表現出更積極的互動行為，值得進行深入的探討。 因此，本研究基於提昇潛水者的社會知覺，於問題導向學習環境中發展「社會互動排名」與「學習夥伴推薦」激勵機制，以探究其對於激發潛水者在社群互動之「討論區與訊息區的文章張貼篇數及內容層次」、「四階段問題導向學習閱讀心得寫作成效」，以及學習社群中的「網路密度」、「網路直徑」、「中心度」。除此之外，也探究「外向－內向」、「人際和諧－人際問題」、「信任感－迫害感」等基本人格特質，是否與潛水者被激發與否的成效有關，進而歸納激發潛水者的具體有效策略。 研究結果顯示，具「社會互動排名」與「學習夥伴推薦」激勵機制之問題導向學習平台，對於提昇社群討論互動以及學習成效具有正向顯著效益；激勵機制確實能有效激發潛水者，降低潛水情形，並且實施激勵機制對於凝聚整體學習社群網絡亦具有正向的效用。 / Lurking is a common behavior in the network community, and lurkers often take the majority in the network community. They often get more from the community, but give less to it. To the whole community, although it doesn’t do any harm, the contribution they make to the network community is so limited, which can’t help the development and growth of the entire community. Therefore, it is quite crucial for the development of the network community about how to motivate the lurkers to participate in the interactive discussion and contribute to the community actively. Especially in the digital learning environment, it should actively develop the strategies to motivate the lurkers effectively, so as to promote the willingness of the lurkers to participate in the interactive discussion of the community more actively. In this way, it can improve the driving force of the cooperative learning in the community. It deserves deep exploration about whether it can help to motivate the lurkers to present more active behaviors in the interaction by making them feel important to participate in the community interaction and improving their social awareness. Therefore, based on the purpose of improving the social awareness of the lurkers, this study develops the motivation mechanism of “social interaction ranking” and ” learning partner recommendation” in the learning-oriented environment to explore the effects of motivating the lurkers in the community interaction, such as “the number and content levels of the articles posted in the forum and bulletin board”, “writing effects of the four-stage problem-based learning and reading”, as well as the “network density”, “network diameter” and “concentration” in the learning community. Besides, it also discusses whether the basic personality is correlated to the effect of motivating the lurkers, including “introversion-extraversion”, “interpersonal problems- interpersonal harmony”, “sense of persecution- sense of trust”, so as to further summarize the concrete and effective strategies of motivating the lurkers. The study results show the problem-based learning platform with the motivation mechanism of “social interaction ranking” and ” learning partner recommendation” show positive and significant benefits to improve the social discussion interaction and learning effect. Moreover, the motivation mechanism system is proven to motivate the lurkers and reduce the lurking situation effectively, and the practice of the motivation mechanism system has positive effect on the cohesion of the whole learning community.

客戶關係管理（Customer Relationship Management, CRM）藉由各種資訊技術來留住客戶，以產生更多的商業價值。然而，許多文獻指出，CRM系統的失敗率很高，尤其是CRM主要的核心能力--銷售員自動化（Sales Force Automation, SFA）。研究指出改善的方式包含更好的管理支援、培訓、系統易用性和強烈的使用動機等等。接續此建議，本文提出了一個銷售員支援（Sales Force Support, SFS）系統，藉由線上分析處理（Online Analytical Processing, OLAP）、資料採礦（Data Mining, DM）和雲端服務（Cloud Service）等技術，協助彙整及提供支援銷售員的客戶推薦 (Customer Recommendation)和自我績效評估(Self Evaluation)功能，以刺激更好的銷售能力、滿足客戶與管理。可望提高系統的易用性和業務人員的使用動機，藉以橋接銷售員和管理人員之間的差異。為了評估推薦功能之適用性，本論文也發展一套驗證指標，並採用一套隨機數學模型（Stochastic Mathematical Model），作為強化推薦預測之嘗試。 / Customer Relationship Management (CRM) adopts various information technologies to retain and attain customers in order to generate more business values. However, the earlier studies indicate the failure rate for CRM systems is high and it’s even higher for Sales Force Automation (SFA), a major core in CRM. They usually suggest the enhancement in better management support, more training, user friendliness, and usage motivation, and so on. Following the suggestions, this research proposes a Sales Force Support (SFS) system to integrate technologies like OLAP (Online Analytical Processing), Data Mining (DM), and cloud service, etc. to provide supporting information in customer recommendation and self-evaluation, in order to better stimulate sales and satisfy customer and management. The objectives can be achieved by enhancing the user friendliness and usage motivation, and bridging the differences between sales force and management. To evaluate the fitness of recommendation function, a set of validation measures is also developed. In addition, a stochastic mathematical model is also attempted to enhance the recommendation prediction.

本研究檢測會計師事務所總所審計與分析師盈餘預測行為的相關性。以中國大陸2010年至2015年A股上市公司為研究對象，構造分析師盈餘預測行為的回歸模型，並以分析師追蹤人數、分析師盈餘預測準確度及預測分歧度三種特性進行分析。 研究結果顯示，會計師事務所總所審計與分析師追蹤人數、盈餘預測準確度皆呈顯著正相關，與預測分歧度呈顯著負相關。進一步檢測發現：總所的審計公費更高，經會計師事務所總所審計的企業，分析師更願意對其股票給予較高的投資評級。這也顯示會計師事務所總所付出的努力更多，審計品質更好，因而分析師對其會計資訊信賴程度更高，對該公司之追蹤意願更高，盈餘預測誤差與預測分歧度更低，分析師也更願意推薦其股票。 / The primary objective of this thesis is to explore the association between headquarter office auditors and analysts’ earnings behaviors. I use a sample of firm observations from China during 2010-2015. The main findings can be summarized as follows. I find that firms audited by headquarter office auditors have more analysts following compared to those audited by branch office auditors. Secondly, I find that analysts’ earnings forecasts are more accurate and less dispersed for firms audited by headquarter office auditors than firms audited by branch office auditors. Further analysis indicates that the reason for the above results is that headquarter office auditors exert more effort, measured as audit fees, than branch office auditors. Finally, the empirical results indicate that analysts make more favorable recommendations for firms audited by headquarter office auditors than for those audited by branch office auditors. Overall, the findings suggest that headquarter office auditors have better audit quality and in turn result in more analysts following and issuing higher-quality forecasts and favorable recommendations.