This work aims at studying the effect of word position in text on understanding and tracking the content of written text. In this thesis we present two uses of word position in text: topic word selectors and topic flow signals. The topic word selectors identify important words, called <i>topic words</i>, by their spread through a text. The underlying assumption here is that words that repeat across the text are likely to be more relevant to the main topic of the text than ones that are concentrated in small segments. Our experiments show that manually selected keywords correspond more closely to topic words extracted using these selectors than to words chosen using more traditional indexing techniques. This correspondence indicates that topic words identify the topical content of the documents more than words selected using the traditional indexing measures that do not utilize word position in text. The second approach to applying word position is through <i>topic flow signals</i>. In this representation, words are replaced by the topics to which they refer. The flow of any one topic can then be traced throughout the document and viewed as a signal that rises when a word relevant to the topic is used and falls when an irrelevant word occurs. To reflect the flow of the topic in larger segments of text we use a simple smoothing technique. The resulting smoothed signals are shown to be correlated to the ideal topic flow signals for the same document. Finally, we characterize documents using the importance of their topic words and the spread of these words in the document. When incorporated into a Support Vector Machine classifier, this representation is shown to drastically reduce the vocabulary size and improve the classifier's performance compared to the traditional word-based, vector space representation.

Receiver operating characteristic (ROC) curves are widely used for evaluating classifier performance, having been applied to e.g. signal detection, medical diagnostics and safety critical systems. They allow examination of the trade-offs between true and false positive rates as misclassification costs are varied. Examination of the resulting graphs and calcu- lation of the area under the ROC curve (AUC) allows assessment of how well a classifier is able to separate two classes and allows selection of an operating point with full knowledge of the available trade-offs. In this thesis a multi-objective evolutionary algorithm (MOEA) is used to find clas- sifiers whose ROC graph locations are Pareto optimal. The Relevance Vector Machine (RVM) is a state-of-the-art classifier that produces sparse Bayesian models, but is unfor- tunately prone to overfitting. Using the MOEA, hyper-parameters for RVM classifiers are set, optimising them not only in terms of true and false positive rates but also a novel measure of RVM complexity, thus encouraging sparseness, and producing approximations to the Pareto front. Several methods for regularising the RVM during the MOEA train- ing process are examined and their performance evaluated on a number of benchmark datasets demonstrating they possess the capability to avoid overfitting whilst producing performance equivalent to that of the maximum likelihood trained RVM. A common task in bioinformatics is to identify genes associated with various genetic conditions by finding those genes useful for classifying a condition against a baseline. Typ- ically, datasets contain large numbers of gene expressions measured in relatively few sub- jects. As a result of the high dimensionality and sparsity of examples, it can be very easy to find classifiers with near perfect training accuracies but which have poor generalisation capability. Additionally, depending on the condition and treatment involved, evaluation over a range of costs will often be desirable. An MOEA is used to identify genes for clas- sification by simultaneously maximising the area under the ROC curve whilst minimising model complexity. This method is illustrated on a number of well-studied datasets and ap- plied to a recent bioinformatics database resulting from the current InChianti population study. Many classifiers produce “hard”, non-probabilistic classifications and are trained to find a single set of parameters, whose values are inevitably uncertain due to limited available training data. In a Bayesian framework it is possible to ameliorate the effects of this parameter uncertainty by averaging over classifiers weighted by their posterior probabil- ity. Unfortunately, the required posterior probability is not readily computed for hard classifiers. In this thesis an Approximate Bayesian Computation Markov Chain Monte Carlo algorithm is used to sample model parameters for a hard classifier using the AUC as a measure of performance. The ability to produce ROC curves close to the Bayes op- timal ROC curve is demonstrated on a synthetic dataset. Due to the large numbers of sampled parametrisations, averaging over them when rapid classification is needed may be impractical and thus methods for producing sparse weightings are investigated.

La documentation des programmes aide les développeurs à mieux comprendre le code source pendant les tâches de maintenance. Toutefois, la documentation n’est pas toujours disponible ou elle peut être de mauvaise qualité. Le recours à la redocumentation s’avère ainsi nécessaire. Dans ce contexte, nous proposons de faire la redocumentation en générant des commentaires par application de techniques de résumé par extraction. Pour mener à bien cette tâche, nous avons commencé par faire une étude empirique pour étudier les aspects quantitatifs et qualitatifs des commentaires. En particulier, nous nous sommes intéressés à l’étude de la distribution des commentaires par rapport aux différents types d’instructions et à la fréquence de documentation de chaque type. Aussi, nous avons proposé une taxonomie de commentaires pour classer les commentaires selon leur contenu et leur qualité. Suite aux résultats de l’étude empirique, nous avons décidé de résumer les classes Java par extraction des commentaires des méthodes/constructeurs. Nous avons défini plusieurs heuristiques pour déterminer les commentaires les plus pertinents à l’extraction. Ensuite, nous avons appliqué ces heuristiques sur les classes Java de trois projets pour en générer les résumés. Enfin, nous avons comparé les résumés produits (les commentaires produits) à des résumés références (les commentaires originaux) en utilisant la métrique ROUGE. / Programs documentation is very useful to programmers during maintenance tasks, especially for program comprehension. However, the documentation is not always available or it may be badly written. In such cases, redocumentation becomes so necessary. In this work, we propose a redocumentation technique that consists in generating comments by using extraction summary techniques. For this purpose, we conducted an empirical study to analyze the quantitave and qualitative aspects of comments. Particularly, we were interested in studying comment distribution over different types of construct and studying the frequency of documentation for each construct type. We propose a comment taxonomy to classify them according to their content and quality. Given the results of the empirical study, we decided to summarize Java classes by extracting the comments of methods and constructors. We defined several heuristics in order to determine the most relevant comments to be extracted. After that, we applied these heuristics to Java classes from three projects in order to generate summaries. Finally, we compared the generated summaries (generated comments) to the reference ones (original comments) by using the metric ROUGE.

Hintergrund: Leptin ist ein unabhängiger kardiovaskulärer Risikofaktor und scheint prothrombotische Effekte zu besitzten. Adiponectin hingegen scheint vor Thrombosen zu schützen. Wir untersuchten deswegen die Bedeutung dieser beiden Adipozytokine bei Patienten mit einer akuten Lungenembolie. Durchführung: Im Rahmen einer prospektiven Fall-Kontroll-Studie schlossen wir zwischen 2003 und 2006 97 Patienten mit gesichterter Lungenembolie und 40 Patienten mit dem Ausschluß einer Lungenembolie ein. Die Lungenembolie-Patienten wurden bezüglich eines komplizierten Verlaufs innerhalb der ersten 30 Tage nach dem Ereignis (Tod, Katecholaminebedarf, Reanimation, Intubation, Hypotonie) sowie bezüglich des Langzeit-Überlebens beobachtet. Ergebnisse: Bei Patienten mit Lungenembolie ist ein höherer Leptinspiegel ein prognostischer Parameter für eine niedrigere 30-Tages-Komplikationsrate und ein besseres Langzeit-Überleben. Patienten mit einer Lungenembolie weisen signifikant höhere Adiponectinwerte als Patienten ohne Lungenembolie auf. Eine prognostische Bedeutung hat Adiponectin bei Patienten mit einer Lungenembolie nicht. Schlußfolgerung: Niedrige Leptinspiegel sind bei Patienten mit einer akuten Lungenembolie ein unabhängiger prognostischer Parameter für einen komplizierten 30-Tages-Verlauf und eine erhöhte Mortalität im Langzeit-Verlauf. Lungenembolie-Patienten weisen eine Hyperoadiponectinämie auf, dies könnten Ausdruck einer kardialen Sekretion und Wirkung des Asiponectins infolge der Lungenembolie sein.

Les ontologies offrent une modélisation des connaissances d'un domaine basée sur une hiérarchie des concepts clefs de ce domaine. Leur utilisation dans le cadre des Systèmes de Recherche d'Information (SRI), tant pour indexer les documents que pour exprimer une requête, permet notamment d'éviter les ambiguïtés du langage naturel qui pénalisent les SRI classiques. Les travaux de cette thèse portent essentiellement sur l'utilisation d'ontologies lors du processus d'appariement durant lequel les SRI ordonnent les documents d'une collection en fonction de leur pertinence par rapport à une requête utilisateur. Nous proposons de calculer cette pertinence à l'aide d'une stratégie d'agrégation de scores élémentaires entre chaque document et chaque concept de la requête. Cette agrégation, simple et intuitive, intègre un modèle de préférences dépendant de l'utilisateur et une mesure de similarité sémantique associée à l'ontologie. L'intérêt majeur de cette approche est qu'elle permet d'expliquer à l'utilisateur pourquoi notre SRI, OBIRS, estime que les documents qu'il a sélectionnés sont pertinents. Nous proposons de renforcer cette justification grâce à une visualisation originale où les résultats sont représentés par des pictogrammes, résumant leurs pertinences élémentaires, puis disposés sur une carte sémantique en fonction de leur pertinence globale. La Recherche d'Information étant un processus itératif, il est nécessaire de permettre à l'utilisateur d'interagir avec le SRI, de comprendre et d'évaluer les résultats et de le guider dans sa reformulation de requête. Nous proposons une stratégie de reformulation de requêtes conceptuelles basée sur la transposition d'une méthode éprouvée dans le cadre de SRI vectoriels. La reformulation devient alors un problème d'optimisation utilisant les retours faits par l'utilisateur sur les premiers résultats proposés comme base d'apprentissage. Nous avons développé une heuristique permettant de s'approcher d'une requête optimale en ne testant qu'un sous-espace des requêtes conceptuelles possibles. Nous montrons que l'identification efficace des concepts de ce sous-espace découle de deux propriétés qu'une grande partie des mesures de similarité sémantique vérifient, et qui suffisent à garantir la connexité du voisinage sémantique d'un concept.Les modèles que nous proposons sont validés tant sur la base de performances obtenues sur des jeux de tests standards, que sur la base de cas d'études impliquant des experts biologistes. / Domain ontologies provide a knowledge model where the main concepts of a domain are organized through hierarchical relationships. In conceptual Information Retrieval Systems (IRS), where they are used to index documents as well as to formulate a query, their use allows to overcome some ambiguities of classical IRSs based on natural language processes.One of the contributions of this study consists in the use of ontologies within IRSs, in particular to assess the relevance of documents with respect to a given query. For this matching process, a simple and intuitive aggregation approach is proposed, that incorporates user dependent preferences model on one hand, and semantic similarity measures attached to a domain ontology on the other hand. This matching strategy allows justifying the relevance of the results to the user. To complete this explanation, semantic maps are built, to help the user to grasp the results at a glance. Documents are displayed as icons that detail their elementary scores. They are organized so that their graphical distance on the map reflects their relevance to a query represented as a probe. As Information Retrieval is an iterative process, it is necessary to involve the users in the control loop of the results relevancy in order to better specify their information needs. Inspired by experienced strategies in vector models, we propose, in the context of conceptual IRS, to formalize ontology based relevance feedback. This strategy consists in searching a conceptual query that optimizes a tradeoff between relevant documents closeness and irrelevant documents remoteness, modeled through an objective function. From a set of concepts of interest, a heuristic is proposed that efficiently builds a near optimal query. This heuristic relies on two simple properties of semantic similarities that are proved to ensure semantic neighborhood connectivity. Hence, only an excerpt of the ontology dag structure is explored during query reformulation.These approaches have been implemented in OBIRS, our ontological based IRS and validated in two ways: automatic assessment based on standard collections of tests, and case studies involving experts from biomedical domain.

Master of Music / Department of Music / Phillip D. Payne / The purpose of this study was to determine if music educators who have had experiences with informal music are more likely to employ informal learning within their classroom. Secondary research objectives included a comparison of survey results against specific demographics of the survey participants, an examination of the types of informal learning that the participants experienced and facilitated, and a look at the perceived barriers and benefits of informal music learning from the viewpoints of the participants. Participants (N=25) were practicing music educators pursuing graduate music studies. The participants were enrolled in a summer Master of Music program at a university in the Midwest. Data was collected by employing a pen and paper survey that provided a demographic description and informal music learning questionnaire. The participants were asked to indicate the frequency of participation in informal music activities prior to becoming a music educator. They further reported what informal music learning activities they facilitate within their school music curriculum. Finally, the participants responded to two short answer questions where they identified barriers and benefits they perceive with the implementation of informal music practices within their music programs. Results from a Pearson correlation showed a moderately strong relationship (p = 0.43) between participants who had informal music experiences (E) and who employed informal music learning within their music curriculum (C). There were no significant differences observed in the results between participants of different gender or school division. Of the short answer responses cataloged, participants cited a lack of experience with informal music and difficulty of connecting informal music learning to the formal music curriculum as the barriers to employing informal music learning in the classroom. The participants discussed the increase in student motivation, expanding musicality, and real-world relevance as the benefits of informal music learning. Knowledge gained from this study may be useful to individuals facilitating informal music learning within music education programs at the primary, secondary, or collegiate levels.

Relevance has become one of the key priorities for accounting policy, and implies that accounting should have an impact on the economic decision-making of accounting users. Despite the increased importance given to users, however, little is known concerning the properties of such relevance in practice. Furthermore, the lack of insight into the practices of users has been mitigated with a theoretical perspective of decision-making which supports an insufficient understanding of how stock markets function and how accounting users behave. This dissertation contributes to the emerging interest in the sociology of financial analysis by following users in their pursuit of relevance. By theorising financial analysis as a social and institutional practice, this dissertation investigates not only how accounting is relevant but also how such relevance is influenced by the particular setting of accounting users. Furthermore, the understanding of relevance as located within the activities leading up to a decision is here extended by emphasising the continuous activities of users and therefore also the role of accounting in the management of their decisions. Based on in-depth field studies targeting the activities of (sell-side) equity research analysts and equity sales brokers, this dissertation presents four papers addressing different notions of accounting, users and relevance. Theoretical insights are drawn from sociology and include actor-network theory, dramaturgy and text-and-conversation-theory. The studies find that the organisation of the sell-side industry necessitates a use of accounting which accentuates the links between accounting, users and investments recommendations. This dissertation concludes that, in order to produce and sustain such links, relevance becomes (a) mediated by a variety of elements, (b) based on the production of differences, and (c) mutually constitutive for accounting and users. / <p>At the time of the doctoral defense, the following papers were unpublished and had a status as follows: Paper 2: Manuscript. Paper 3: Manuscript. Paper 4: Manuscript.</p><p> </p>

Philosophiae Doctor - PhD / Since a number of studies on textbooks in Second Language Acquisition (SLA) have shown that textbooks hold a major place in education (De Guzman, 2000; Oakes & Saunders, 2004), this study proposes to investigate the content of an English Language Teaching (ELT)/ English as a Second Language (ESL) textbook used in 6th form secondary school (Grade 12) in the DRC, Go for English 1RE. It aims to identify and evaluate the content of this ELT textbook so as to deduce and derive main insights for the determination or not of its appropriateness and relevance in terms of its contribution to language use and literacy in the ELT/ESL curriculum of the DRC. The study draws its theoretical underpinning from two theories: the Cunningsworth’s textbook analysis theory and McDonough and Shaw evaluation theory. As research design, the study adopts an a descriptive, exploratory and interpretive design which draws on both quantitative and qualitative data collected on the basis of textbook evaluation checklist and semi-structured interviews. In regard to the procedural orientation, the study uses descriptive and content analysis to analyze, interpret and examine both interviews and textbook evaluation likert-scale checklist data. In respect of its data, the study uses ‘mixed methods approach’. Both qualitative and quantitative data come from 259 teacher and student participants on the basis of two different samples. The quantitative data comes from 209 student participants and 25 teacher participants and the qualitative data from 10 student participants and 15 teacher participants. The findings attest to the general content of "Go for English 1RE ELT" textbook in regard to language activities and tasks related to its subject matter, to the quality and nature of language it contains, and finally to the diversity in its subject matter and its cultural aspects, is suitable for language use and literacy skills development. However, even though its content is suitable, the findings also indicate that this ELT textbook is not well adapted to Congolese 6th form secondary school students’ level.

Abstract Background “Traditional digital advertising has become wallpaper. It doesn’t improve anyone’s experience on a site and readers, myself included, pretty much look past it” stated by Jason Hill, Global Head of Media Strategies (Adage.com, 2019). Marketing managers attempt and fail in adapting traditional marketing strategies to the unfamiliar digital landscape. Traditional techniques e.g. banner ads appear as intrusive and unwanted according to consumers. Brands have in response strategically changed its communication to a rather subtle approach that aims to be non-intrusive. Aforesaid strategy is referred to as native advertising. Native advertising however is not unproblematic as said approach may be exploited to be deceptive, intrusive and secretive, brand whom implements such a strategy may be rewarded with short term positives. However the long term effects of said strategy prove negatives in various aspects of the brand. Purpose The purpose of this thesis is to explain how the conceptualization of native ads influence users’ feeling of non-intrusiveness. Method The conducted study implemented a deductive approach to research, including an explanatory purpose as previously elaborated upon. The study is further characterized as a cross-sectional design in which the researchers aimed to test the conceptualization of non-intrusive native advertising. The researchers further conducted a self administrated questionnaire with a total of 158 respondents. The data was cleaned in SPSS and further tested for reliability, validity and hypothesis testing. Conclusion The findings of the conducted study concludes that personalized advertising positively contributes to users’ feeling of non-intrusiveness. On the other hand, contextual relevance, 1 transparency, emotional appeal and informational appeal were rejected as such variables were not statistically significant in explaining non-intrusiveness. Therefore, the researchers rejected hypothesis H1, H2, H3, H4 and therefore rejected previous research stating that these variables contributes to users’ feeling of non-intrusiveness. However H5 was accepted, therefore theories that suggests that personalization according to one’s interest, previous interactions and real-time data to influence users feeling of non-intrusiveness were confirmed.

Verificada a existência de diversos entendimentos conflitantes a respeito da relevância que a boa ou a má-fé desempenham para a configuração da fraude de execução e dos riscos decorrentes da indefinição quanto à matéria, propõe-se, nesta tese, a apresentação de nova hipótese interpretativa que, à luz da história e do regramento atual do instituto, identifique qual é e em que casos se faz relevante o elemento subjetivo na fraude de execução. Para atingir esse objetivo, parte-se da análise das hipóteses de fraude de execução previstas na legislação em vigor e do estabelecimento de algumas premissas conceituais necessárias ao desenvolvimento do trabalho. Também no início da tese, é exposto o entendimento atualmente dominante na doutrina e na jurisprudência em relação à matéria, apresentando-se críticas em relação ao mesmo, com base em pesquisa de casos concretos julgados pelo Superior Tribunal de Justiça. Em seguida, são identificadas a feição e a relevância do elemento subjetivo tal como originalmente apresentadas na história do instituto, confrontando-as com as que lhe estabelece entendimento dominante e demonstrando a sua prevalência tanto à luz da legislação em vigor, quanto em relação à moderna concepção da responsabilidade patrimonial. A tese prossegue com a apresentação de sugestão de metodologia para análise dos casos de fraude de execução. Finalmente, conclui-se o trabalho com a verificação das propostas legislativas para a disciplina futura da fraude de execução, apresentando-se sugestão alternativa, baseada na análise dos problemas verificados e na solução oferecida pela hipótese apresentada nesta tese. / Once verified the existence of many conflicting positions about the relevance that the good or bad faith may have on the configuration of the fraude de execução (literally: fraud of execution) and the risks deriving from the lack of definition about this subject, we propose, on this thesis, to present a new interpretative hypothesis which, on the light of the history and recent ruling of the institute, identifies what is the subjective element in the fraude de execução and in which cases it becomes relevant. In order to achieve such aim, we start from the analysis of the hypothesis of fraude de execução foreseen in the current law and by establishing some conceptual premises that are necessary to the development of the work. Also, in the beginning of the thesis, we expose the prevailing position on the current jurisprudence and doctrine and criticize it based on the research of precedents of our Superior Court (Superior Tribunal de Justiça). Afterwards, we identify the features and relevance of the subjective element such as originally presented on the history of the institute, comparing them with those established by the dominant position and demonstrating its prevalence, even on the light of the effective rules, than in relation to the modern concept of patrimonial responsibility. The thesis continues with the suggestion of a methodology to analyze the cases of fraude de execução. Finally, the work is concluded with the study of the bills regarding the future ruling of the fraude de execução and an alternative suggestion is presented, based on the analysis of the problems verified and on the solution offered by the hypothesis presented on this thesis.