Doctor of Philosophy / Department of Computer Science / William H. Hsu / Current research in the field of recommender systems takes into consideration the interaction between users and items; we call this the homogeneous setting. In most real world systems, however these interactions are heterogeneous, i.e., apart from users and items there are other types of entities present within the system, and the interaction between the users and items occurs in multiple contexts and scenarios. The presence of multiple types of entities within a heterogeneous information network, opens up new interaction modalities for generating recommendations to the users. The key contribution of the proposed dissertation is representation learning in heterogeneous information networks for the recommendations task. Query-based information retrieval is one of the primary ways in which meaningful nuggets of information is retrieved from large amounts of data. Here the query is represented as a user's information need. In a homogeneous setting, in the absence of type and contextual side information, the retrieval context for a user boils down to the user's preferences over observed items. In a heterogeneous setting, information regarding entity types and preference context is available. Thus query-based contextual recommendations are possible in a heterogeneous network. The contextual query could be type-based (e.g., directors, actors, movies, books etc.) or value-based (e.g., based on tag values, genre values such as ``Comedy", ``Romance") or a combination of Types and Values. Exemplar-based information retrieval is another technique for of filtering information, where the objective is to retrieve similar entities based on a set of examples. This dissertation proposes approaches for recommendation tasks in heterogeneous networks, based on these retrieval mechanisms present in traditional information retrieval domain.

No description available.

金融市場中存在著許多預測機構, 他們各自召集信眾並且不時地釋放訊息給其會員好讓他們能在交易中獲利。每筆訊息皆代表著各機構對此資產價值的預測, 會員則依此訊息至市場上尋找機會交易。他們交易前會先理性地觀察市場過往的波動。如果市場走勢所預測的訊息與自己的訊息一致, 那交易者交易時大概不會有所顧慮。然而當市場趨勢與自己的訊息不一致時, 交易者勢必會陷入兩難。仔細地衡量斟酌兩股力量的輕重後, 進而選擇他覺得對的決定。如果交易者放棄自己的訊息而追隨前人交易的腳步, 那我們可定義這是一種群聚的行為。 如果某一機構的會員人數龐大, 則他們勢必會影響市場價格的波動。不知情的交易者在看到價格趨勢如此時, 可能會放棄自己的訊息轉而追隨過往交易者的選擇。然而此種交易伴隨著風險, 因為不知道正確的訊息為何, 當價格已經達到機構所預測的目標時, 知情的會員便開始反向操作, 而不知情的交易者可能會持續地採取此一交易策略。於是當資產真正價值揭露時, 不知情交易者便可能因此被套牢。跟隨大眾的決策相對保險, 但是當追隨的人沒有額外的訊息無法查覺情勢的變化時, 便可能面臨損失的風險。 我們建構了一個存在兩種類型交易者的市場, 一方是沒有參加機構的一般交易者, 另一方是同時參加某一機構的會員交易者。透過私有訊息與公開歷史交易預測的權衡, 交易者必須想辦法在這一次機會的交易中獲利。而我們想找出是否對任何交易者而言, 參加預測機構是有利可圖的。 想當然爾, 市場中會員交易者的多寡對於機構預測目標價位的達成頗為重要, 因為影響力的大小間接決定了市場雙方的利潤。當然每位交易者對訊息的信心也有所不同, 這些因素都會影響雙方的利潤。而本篇論文即是嘗試找出在哪種條件之下, 參與機構交易者的交易績效會比沒有參與機構交易者的績效為佳。 / Traders with their own heterogeneous hidden information are coming to the market to trade in order to maximize their expected profits. They will observe the trends of prices and compare it to their private signals and then make the right decisions. The trends might not consistent with the private signals. If the traders choose to abandon his own signals and follow the actions made by predecessors, we called the action “Herds.” In this paper, we set a mechanism to harmonize with these two powers. Also we put the traders into two subgroups, and one of the groups will send another signal to its members. For simplicity, we use a sequential trading model to see the trade patterns. Since we use the closing price to measure traders’ profits, traders in the market need to presume what the closing price will be. Then we calculate the profits of each group and find out their performance. We want to see under what kind of conditions, the performance of one group will be better than that of another group. If we can find the conditions of better performance, it is worth for the traders to join that group.

Most real systems can be represented as a graph of multi-typed components with a large number of interactions. Heterogeneous Information Networks (HIN) are interconnected structures with data of multiple types which support the rich semantic meaning of structural types of nodes and edges. In HIN, different information can be presented using different types and forms of data, but may have the same or complementary information. So there is knowledge to be discovered. Terminology Knowledge Structures (TKS) como terminology products can be sources of linguistic representations and knowledge to be used for enrich the HIN and create a measure of similarity to extract the documents similar to each other, even if these documents are of different types (for example, finding medical articles that are in some way related to medical records). In this sense, this work presents the creation of a Heterogeneous Information Network using classical similarity measures, terminology products and the attributes of documents by an algorithm called NetworkCreator. As a contribution, an algorithm called NetworkCreator was created that from medical records and scientific articles builds an HIN with related documents, was also created. The algorithm HeteSimTKSQuery to calculate similarity measures between documents of different types which are in HIN. Terminology products with meta-paths were also explored. The results were efficient, reaching on average 89\\% accuracy in some cases. However, it is important to note that all HIN presented in the researched literature were constructed only by one type of data coming from a single source. The results show that the algorithms are feasible to solve the problems of HIN construction and search for similarity. But it still needs improvement. In the future one can work on detection in the detection of node granularity of these networks and try to reduce the network construction runtime / A maioria dos sistemas reais pode ser representada como um grafo de componentes multi-tipados com um grande número de interações. Redes de Informação Heterogênea (HIN) são estruturas interconectadas com dados de múltiplos tipos que suportam o rico significado semântico de tipos estruturais de nós e arestas. Nas HIN, diferentes informações podem ser apresentadas usando diferentes tipos e formas de dados, mas podem ter informações iguais ou complementares. Então, há conhecimento a ser descoberto. Estruturas de Conhecimento Terminológicos (TKS) como produtos terminológicos podem ser fontes de representações linguísticas e de conhecimento a ser usado para enriquecer a HIN e criar uma medida de similaridade para extrair os documentos similares entre si, mesmo que esses documentos sejam de tipos diferentes (por exemplo, encontrar os artigos médicos que de alguma forma estão relacionados com registros médicos). Nesse sentido, este trabalho apresenta o algoritmo NetworkCreator que cria uma Rede de Informações Heterogêneas utilizando medidas de similaridade clássicas, produtos de terminológicos e os atributos dos documentos. Nos experimentos, foram utilizados prontuários médicos e artigos científicos para construir a HIN e relacionar seus conteúdos. O algoritmo HeteSimTKSQuery também foi criado para calcular medidas de similaridade entre os documentos de diferentes tipos que se encontram na HIN. Produtos terminológicos com meta-caminhos também foram explorados. Os resultados se mostraram eficientes, alcançando em média 89\\% de acurácia, em alguns casos. No entanto, é importante notar que todas as HIN apresentadas na literatura pesquisada foram construídas apenas por um tipo de dados proveniente de uma única fonte. Os resultados mostram que os algoritmos são viáveis para resolver os problemas de construção de HIN e busca de similaridade. Porém, eles ainda precisam de aperfeiçoamentos. Futuramente, pode-se trabalhar na detecção da granularidade dos nós destas redes e tentar reduzir o tempo de construção da rede

No description available.

隨著資訊科技的發展，會計資訊變得更具即時性與可及性。會計資訊的提供者可利用自動化系統即時產出財務報告，並進行線上財務報導；而會計資訊的需求者，也可利用網際網路不受時空限制的特性，擷取所需的資訊。會計資訊的及時性品質獲得重大改善，而針對異質性的資訊需求者，用更低成本的方式滿足其多元需求的可能性也大為增加。XBRL可說是近年來會計與資訊科技結合的一項重要進展。本論文將以我國會計資訊市場為背景，採用XBRL及其他相關資訊技術，開發出一套能提供異質性會計資訊使用者閱讀及查詢XBRL格式文件之資訊系統。 本研究以圍繞標準財務報導之客製化（Customization Around Standard Report, CASR）架構為藍本，並與客製化報導（Customized Financial Reporting, CFR）理論結合，完成以XBRL為資料格式的客製化財務報導系統（Customized Financial Reporting System, CFRS），俾驗證XBRL在我國會計資訊市場之技術可行性，並為異質性的資訊使用者創造附加價值。在外觀客製化方面，係透過客製化樣版的設計而達成；內容客製化的部分，則因採用XBRL資料格式及CFRS的系統功能，而實現並延伸了部分的內容客製化面向。 / Accounting information has become real-time and more accessible with the development of information technology. Providers of accounting information can use automatic systems to produce real-time financial reports and distribute them on the Internet; and users of accounting information can extract data freely from WWW (World Wide Web). Consequently, the timeliness of accounting information is greatly improved, and heterogeneous information users’ various demands could be satisfied. XBRL is an important achievement on the integration of accounting and information technology in the past few years. This paper uses XBRL and other information techniques to develop an XBRL-based information system for the heterogeneous users of Taiwan accounting information market to read and inquire XBRL documents. Our research is based on the architecture of the Customization Around Standard Report (CASR) model under the Customized Financial Reporting (CFR) theory. We have developed the Customized Financial Reporting System (CFRS) to verify the technical feasibility of XBRL in Taiwan accounting information market, and to create values for heterogeneous information users. We use customized style sheets for the customized presentation of accounting information, and implement several dimensions of content customization by using XBRL and CFRS’ s functionalities.

由於對資訊科技以及網際網路/和企業內網路的依賴持續加深，異質資訊整合在電子化企業中已經成為一個普遍存在而且相當重要的議題。因為在缺乏整合的情形下個別地存取異質資訊來源可能會造成資訊的混亂，而且在電子化企業的環境中，這麼做也不符合成本效益決策支援管理分析。在傳統異質資訊整合的研究中，通常會創造一個共同資料模式來處理異質性的問題，而可延伸性標記語言已經成為網路上交換資訊時的標準文件格式，使得XML成為整合工作中共同資料模式的一個很好的候選者；然而，XML僅能夠處理結構異質性，無法處理語意異質性，而本體論被視為是一個重要而且自然的工具可以用來表現真實世界中模糊不清的語意和關係，因此，在本研究中也加入了本體論以期達到異質資訊整合中的語意互動性。 在本篇論文中，我們提出一個以學名結構導向非特殊隨機式對應的方法來產生全區域綱要方法（Global Schema），以促成非傳統而是以網路為基礎的異質資訊整合。我們也提出一個對異質資訊來源較具智慧性的查詢方法，該查詢方法應用了global-as-view （GAV）全區域景觀導向方法加上本體論觀念運用，可以同時提高對底層異質資訊來源的結構互動性和語意互動性。我們透過雛型系統的實作來驗證本研究所提供的異質資訊整合方法的可行性。 / The research issues of heterogeneous information integration have become ubiquitous and critically important in e-business (EB) with the increasing dependence on Internet/Intranet and information technology (IT). Accessing the heterogeneous information sources separately without integration may lead to the chaos of information requested. It is also not cost-effective in EB settings. A common general way to deal with heterogeneity problems in traditional HII is to create a common data model. The eXtensible Markup Language (XML) has been the standard data document format for exchanging information on the Web. XML only deals with the structural heterogeneity; it can barely handle the semantic heterogeneity. Ontologies are regarded as an important and natural means to represent the implicit semantics and relationships in the real world. And they are used to assist to reach semantic interoperability in HII in this research. In this thesis, we provide a generic construct orientation no ad hoc method to generate the global schema to enable the web-based alternative to traditional HII. We provide a wiser query method over multiple heterogeneous information sources by applying global-as-view (GAV) approach with the use of ontology to enhance both structural and semantic interoperability of the underlying heterogeneous information sources. We construct a prototype implementing the method to provide a proof on the validity and feasibility.

Cette thèse contribue au débat sur les politiques de ciblage d’inflation et de transparence des banques centrales en présentant notamment trois essais théoriques et empiriques sur le sujet. Dans le premier essai, nous étudions théoriquement les conditions sous lesquelles il serait optimal pour une banque centrale d’adopter explicitement un régime de ciblage d’inflation. Nous proposons un nouveau cadre théorique qui combine les deux principales raisons avancées dans la littérature pour expliquer les effets réels à court terme de la politique monétaire et qui sont d’une part, la présence d’informations hétérogènes entre les agents économiques (Phelps, 1970 ; Lucas, 1972), et d’autre part, la rigidité des salaires ou des prix (Taylor, 1980 ; Calvo, 1983). Nous analysons ensuite notre problématique dans ce nouveau cadre en considérant l’interaction entre le degré de rigidité des prix, et le degré de complémentarités stratégiques dans la fixation de prix des firmes. Nos résultats montrent que l’adoption d’un régime de ciblage d’inflation dépend fortement de l’importance relative des paramètres du modèle. En particulier, nous montrons que le ciblage d’inflation devrait être toujours adopté lorsque les complémentarités stratégiques sont faibles, alors que dans le cas contraire, il est optimal uniquement lorsque les prix sont assez rigides et que la banque centrale détient des informations suffisamment précises sur les fondamentaux de l’économie. Dans le second essai, nous utilisons la macroéconomie expérimentale afin d’évaluer dans quelle mesure l’annonce de la cible d’inflation est pertinente dans un cadre de ciblage de l’inflation. Nos résultats montrent que lorsque la banque centrale ne se soucie que de la stabilisation de l’inflation, l’annonce de la cible d’inflation n’apporte pas de gain supplémentaire en termes de performances macro-économiques, par rapport à une politique monétaire active (type règle de Taylor). Cependant, si la banque centrale intègre également la stabilisation de l’activité économique dans ses objectifs, la communication de la cible contribue à réduire la volatilité de l’inflation, du taux d’intérêt, et de l’écart de production, bien que leurs niveaux moyens ne soient pas affectés. Ce résultat fournit ainsi une justification pour l’adoption d’un régime de ciblage flexible d’inflation par la majorité des pays ciblant l’inflation. Enfin dans le troisième essai, nous appliquons une analyse transversale ainsi que la technique des variables instrumentales, afin d’analyser les effets de la transparence des banques centrales sur les résultats macroéconomiques dans les pays émergents. Nous construisons un nouvel indice de transparence qui combine certains aspects de l’indice de transparence globale d’Eijffinger et Geraats (2006), avec ceux de l’indice de transparence sur le comité de politique monétaire de Hayo et Mazhar (2011). Nous analysons ensuite le rôle individuel de chaque composante du nouvel indice en termes de réduction du niveau de l’inflation et de sa volatilité, ainsi que de la volatilité du produit. Contrairement à la littérature antérieure, nous trouvons que le nouvel indice de transparence ainsi que ses aspects économique, politique, procédurale et de transparence sur la politique monétaire impactent négativement le niveau moyen de l’inflation, mais pas sa volatilité dans ces pays. L’unique composante du nouvel indice qui permet de réduire à la fois la volatilité de l’inflation et celle de la production est la transparence opérationnelle. Ces résultats s’avèrent robustes aux différentes spécifications de modèles économétriques utilisés dans cet essai. / This dissertation contributes to the debate on inflation targeting and central bantransparency by presenting three theoretical and empirical essays on the topic. In the first essay, we theoretically investigate the conditions under which it would be optimal for a central bank to explicitly adopt an inflation targeting regime. We propose a new theoretical framework that combines the two main frictions put forward in the literature to explain the real short run effects of monetary policy that is, heterogeneous information among agents (Phelps, 1970; Lucas, 1972), and wage or price rigidities (Taylor, 1980; Calvo, 1983). We then analyze our issue in this new framework by considering the interaction between the degree of price stickiness, and the degree of strategic complementarities in firms’ price setting. Our results show that adopting an inflation targeting regime crucially depends on the relative importance of the model’s parameters. In particular, we show that inflation targeting should always be adopted when strategic complementarities are low, while in the opposite case, it is optimal only if prices are sticky enough and the central bank holds sufficiently accurate information on the fundamentals of the economy. In the second essay, we use experimental macroeconomics to evaluate to what extent communication of the inflation target is relevant in an inflation targeting framework. Our results show that first, when the central bank only cares about inflation stabilization, announcing the inflation target does not make a difference in terms of macroeconomic performance compared to a standard active monetary policy. However, if the central bank also cares about the stabilization of the economic activity, communicating the target helps to reduce the volatility of inflation, interest rate, and output gap although their average levels are not affected. This finding provides a rationale for the adoption of flexible inflation targeting by the majority of inflation targeting countries. In the third essay, using a cross-sectional analysis and instrumental variables technique, we analyze the impact of central bank transparency on macroeconomic outcomes in emerging economies. We build a new index of transparency that combines some aspects of the overall Eijffinger and Geraats (2006) transparency index, with those of monetary policy committee transparency developed in Hayo and Mazhar (2011). We then analyze the individual role of each component of the new index in mitigating inflation and its volatility, as well as output volatility. By contrast to the previous literature, we interestingly find that the overall new index of transparency as well as its political, economic, procedural, and policy aspects negatively impact the average level of inflation, but not its volatility in these countries. The unique component of the new index that reduces the volatility of both inflation and output is operational transparency, and these results are robust to different econometric and instruments setting specifications.

隨著網際網路和企業內部網路的盛行，異質資訊整合成為電子化企業中一個重要的議題，在網路上進行異質資訊整合涉及許多不同新的資訊技術，目前已經有些研究試圖利用延伸標記語言以及本體論當作中介技術來整合異質資訊，為了有效管理企業內的資訊，我們需要一個績效評估模型來衡量異質資訊整合的效能。在本研究中，我們提出了一個在異質資訊整合中運用延伸標記語言及本體論的績效評估工作量模型，並且建立了一個工作量產生器雛形；本研究的目標是希望發展出一個結合延伸標記語言及本體論的工作量模型，以測試在電子化企業中的異質資訊整合是否能整合不同的資訊模型，並且從這些資訊模型中衍生出語意，此工作量模型包含了延伸標記語言與本體論的資料模型與查詢模型，它們是依照延伸標記語言與本體論學名式的資料結構與查詢功能所制訂的，此外，控制模型則定義了績效評估執行環境中所需設定的變數，為了讓此工作量模型能具可攜性和延展性，以便輕易地應用在不同的領域情境中，本研究採取學名結構式且使用者定義、領域獨立的設計方法，最後，我們利用雛形實作來驗證本研究所提出的研究方法。 / With the popularity of Internet/Intranet, heterogeneous information integration becomes a hot IT topic in electronic business (EB) field. Heterogeneous information integration on the Web involves a number of new techniques. There have been research projects applying XML and ontology as mediated techniques to consolidate heterogeneous information. In order to manage and use information more effectively within the enterprise, a benchmark used to evaluate the mechanism of heterogeneous information integration is needed. In this research, we develop a XML and ontology benchmark workload model in heterogeneous information integration, and build a workload generation prototype. The objective of this research is to develop a workload model combines XML and ontology to test whether the heterogeneous information integration system under EB environment can overcome the diverse formats of content and derive meaning from this content. The workload model consists of XML and ontology data model and query model according to the generic data structure and query functionality. Also, a control model is created to set up the benchmark environment. In order to apply the workload model to different scenarios easier, this workload model is designed to be domain independent and generic-construct-based. Finally, we validate the research model through the prototype implementation.

Master of Philosophy (MPhil) / Informationist Science Fiction theory provides a way of analysing science fiction texts and narratives in order to demonstrate on an informational basis the uniqueness of science fiction proper as a mode of fiction writing. The theoretical framework presented can be applied to all types of written texts, including non-fictional texts. In "Informationist Science Fiction Theory and Informationist Science Fiction" the author applies the theoretical framework and its specific methods and principles to various contemporary science fiction works, including works by William Gibson, Neal Stephenson and Vernor Vinge. The theoretical framework introduces a new informational theoretic re-framing of existing science fiction literary theoretic posits such as Darko Suvin's novum, the mega-text as conceived of by Damien Broderick, and the work of Samuel R Delany in investigating the subjunctive mood in SF. An informational aesthetics of SF proper is established, and the influence of analytic philosophy - especially modal logic - is investigated. The materialist foundations of the metaphysical outlook of SF proper is investigated with a view to elucidating the importance of the relationship between scientific materialism and SF. SF is presented as The Fiction of Veridical, Counterfactual and Heterogeneous Information.