[ES] La investigación llevada a cabo en esta tesis doctoral tiene como objetivo general analizar la importancia del poder transformador de la analítica de big data, a través de las capacidades analíticas de big data en el ecosistema español de las pequeñas y medianas empresas. El contexto de la transformación digital ha remodelado la forma de hacer negocios en las organizaciones debido a la complejidad e incertidumbre del entorno, el surgimiento de empresas nativas digitales, la introducción de nuevas tecnologías e industria 4.0 y el aumento de la competitividad de los mercados. Si bien la implantación de la tecnología e infraestructuras digitales ha sido un tema estudiado en la literatura académica en los últimos años, se divisan retos importantes a nivel humano debido a la modificación del contexto laboral, el liderazgo y las habilidades y competencias necesarias para competir de forma exitosa actualmente. Las personas, y no la tecnología, son el centro de la transformación digital y todo cambio organizativo derivado debe priorizarlas. Muchas de las empresas que invierten en tecnologías como el big data son incapaces de extraer el valor que éste puede ofrecer a través de la analítica de datos, y por tanto, no lo utilizan para tomar decisiones de valor para la organización que lleven a un incremento del desempeño. Tienen poco desarrolladas las capacidades analíticas de big data, necesarias para aprovechar la transformación digital y promover una implementación efectiva de las nuevas tecnologías. De esta problemática derivamos la importancia de conocer cuáles son los antecedentes de las capacidades analíticas de big data y su efecto en ellas, con el objetivo de conseguir un verdadero impacto en el desempeño organizativo. Por una parte, la cultura organizativa se ha identificado como una de las barreras al cambio o un factor impulsor que permite efectuar una transformación digital efectiva. Para ello es necesario implantar nuevas formas de trabajar y adquirir habilidades y conocimientos adecuados que permitan tomar decisiones en base al análisis de los datos. Es por tanto fundamental, que la cultura organizativa promueva e incentive la promoción de capacidades analíticas de big data y la transformación digital. Por otra parte, se destaca el papel del CEO de la organización, y de su visión estratégica orientada al dato para incentivar, liderar y motivar el cambio hacia una transformación digital. El rol del directivo es crucial para motivar un cambio cultural que permita ver la transformación digital y las capacidades analíticas de big data como instrumentos para lograr una mejora de la competitividad, desempeño, creación de valor y aumento de la reputación y satisfacción de las personas. Por tanto, el CEO debe tener un compromiso con la transformación digital tangible y visión estratégica orientada a los datos para tomar decisiones y planificar la estrategia a seguir por toda la organización. Entre las conclusiones del estudio se destaca en primer lugar la relación positiva y significativa de las capacidades analíticas de big data con la transformación digital y el desempeño organizativo a través de la innovación. Además, se pone en valor la importancia de la cultura organizativa y de la orientación al dato, así como de un nivel adecuado de madurez digital, como antecedentes de las capacidades analíticas de big data. Finalmente se analizan los diversos arquetipos culturales para destacar que una cultura digital, jerárquica o adhocrática favorecen la creación de capacidades analíticas y por tanto incrementan el proceso de transformación digital. A partir de las conclusiones se deriva la necesidad de inversión en formación para las personas en capacidades digitales y analíticas y el rol clave del directivo para conseguir una transformación digital exitosa y aprovechar la inversión tecnológica realizada. Por último, se destaca la importancia del diagnóstico cultural y elaboración de un plan de cambio cultural. / [CA] La investigació duta a terme en aquesta tesi doctoral té com a objectiu general analitzar la importància del poder transformador de l'analítica de big data, a través de les capacitats analítiques de big data en l'ecosistema espanyol de les petites i mitjanes empreses. El context de la transformació digital ha remodelat la manera de fer negocis en les organitzacions a causa de la complexitat i incertesa de l'entorn, el sorgiment d'empreses natives digitals, la introducció de noves tecnologies i indústria 4.0 i l'augment de la competitivitat dels mercats. Tot i que la implantació de la tecnologia i infraestructures digitals ha sigut un tema estudiat en la literatura acadèmica en els últims anys, s'albiren reptes importants a nivell humà a causa de la modificació del context laboral, el lideratge i les habilitats necessàries per a competir de manera exitosa actualment. Les persones, i no la tecnologia, són el centre de la transformació digital i tot canvi organitzatiu derivat ha de prioritzar-les. Moltes de les empreses que inverteixen en tecnologies com el big data són incapaços d'extraure el valor que aquest pot oferir a través de l'analítica de dades, i per tant, no l'utilitzen per a prendre decisions de valor per a l'organització que porten a una millora dels resultats. La raó és que tenen poc desenvolupades les capacitats analítiques de big data, necessàries per a aprofitar la transformació digital i promoure una implementació efectiva de les noves tecnologies. D'aquesta problemàtica derivem la importància de conéixer quins són els antecedents de les capacitats analítiques de big data i el seu efecte en elles, amb l'objectiu d'aconseguir un vertader impacte en la millora dels resultats. D'una banda, la cultura organitzativa s'ha identificat com una de les barreres al canvi per a efectuar una transformació digital efectiva. Per aquesta raó cal implantar noves maneres de treballar i adquirir habilitats i coneixements adequats que permeten prendre decisions sobre la base de l'anàlisi de les dades. És per tant fonamental, que la cultura organitzativa promoga i incentive la promoció de capacitats analítiques de big data i la transformació digital. D'altra banda, es destaca el paper del CEO de l'organització, i de la seua visió estratègica orientada a les dades per a incentivar, liderar i motivar el canvi cap a una transformació digital. El paper directiu és crucial per a motivar un canvi cultural que permeta veure la transformació digital i les capacitats analítiques de big data com a instruments per a aconseguir una millora de la competitivitat, acompliment, creació de valor i augment de la reputació i satisfacció de les persones. Per tant, el CEO ha de tindre un compromís amb la transformació digital tangible i una visió orientada a les dades per a prendre decisions i planificar l'estratègia a seguir per tota l'organització. Entre les conclusions de l'estudi es destaca en primer lloc la relació positiva i significativa de les capacitats analítiques de big data amb la transformació digital i la millora dels resultats a través de la innovació. A més, es posa en valor la importància de la cultura organitzativa i de l'orientació a la dada, així com d'un nivell adequat de maduresa digital, com a antecedents de les capacitats analítiques de big data. Finalment s'analitzen els diversos arquetips culturals i es destaca que una cultura digital, jeràrquica o adhocrática afavoreixen la creació de capacitats analítiques i per tant incrementen l'éxit del procés de transformació digital. A partir de les conclusions es deriven algunes implicacions pràctiques com la necessitat d'inversió en formació per a les persones en competències i capacitats digitals i analítiques, el rol clau del directiu per a aconseguir una transformació digital exitosa i aprofitar la inversió tecnològica. Finalment es destaca la importància del diagnòstic cultural i elaboració d'un pla de canvi cultural alineat amb els objectius envers la transformació digital. / [EN] The general objective of the research carried out in this doctoral thesis is to analyse the importance of the transformative power of big data analytics through big data analytical capabilities in the Spanish context of small and medium-sized enterprises. The context of digital transformation has reshaped the way of doing business in organisations due to the complexity and uncertainty of the environment, the emergence of digital native companies, the introduction of new technologies and the increased competitiveness of markets. Whilst the implementation of technology and digital infrastructures has been covered in the academic literature in recent years, there are significant challenges at the human level due to the changing context of work, leadership and the skills needed to compete successfully today. People, not technology, are at the heart of digital transformation and any resulting organisational change must priorise them. Many companies that invest in technologies such as big data are unable to extract the value that big data can offer through data analytics, and therefore do not use it to make valuable decisions for the organisation that lead to increased performance. They have underdeveloped big data analytical capabilities, which are necessary to take advantage of digital transformation and promote the effective implementation of new technologies. From this problem the importance of knowing the background of big data analytical capabilities and their effect on them is derived, in order to achieve a real impact on organisational performance. On the one hand, organisational culture has been identified as a barrier or booster of change for an effective digital transformation. This requires the implementation of new ways of working and the acquisition of appropriate skills and knowledge to enable data-driven decision making. It is therefore essential that the organisational culture promotes and encourages the promotion of big data analytical capabilities and digital transformation. On the other hand, the role of the CEO of the organisation, and his or her data-driven strategic vision to incentivise, lead and motivate change towards digital transformation is highlighted. The role of the top management is crucial to motivate a cultural change that allows to see digital transformation and big data analytics capabilities as instruments to achieve superior outcomes (i.e., improved competitiveness, performance, value creation and increased reputation and people satisfaction). Therefore, the CEO must have a strong commitment to digital transformation and a data-driven orientation to make decisions and settle the strategy for the entire organisation. Among the conclusions of the study, the positive and significant relationship of big data analytics capabilities with digital transformation and organisational performance through innovation are highlighted. This thesis points out the importance of organisational culture and data orientation, as well as an appropriate level of digital maturity, as antecedents to big data analytics capabilities. Finally, the various cultural archetypes are analysed to highlight that a digital, hierarchical or adhocratic culture favours the creation of analytical capabilities and therefore enhances the digital transformation process. From the conclusions, some practical implications are derived, such as the need to invest in training people in digital and analytical skills and capabilities, the key role of the manager in achieving a successful digital transformation and leveraging technological investment. Finally, the importance of cultural diagnosis and the development of a cultural change plan aligned with the strategic objectives for digital transformation is highlighted, and practical recommendations are settled. / Tesis elaborada gracias al apoyo de las Ayudas de Formación del Profesorado Universitario (FPU) otorgadas por el Ministerio de Educación, Cultura y Universidades, del Gobierno de España, y a la Cátedra de Empresa y Humanismo de la Universitat de València. / Orero Blat, M. (2023). Enhancing Big Data Analytics Capabilities: The Influence of Organisational Culture and Data-Driven Orientation [Tesis doctoral]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/191788

Land surface models (LSMs) implemented in numerical weather prediction (NWP) models use several parameters to suitably describe the surface and its interaction with the atmosphere, whose determination is often affected by many uncertainties, strongly influencing simulation results. However, the sensitivity of meteorological model results to these parameters has not yet been studied systematically, especially in complex terrain, where uncertainty is expected to be even larger. This work aims at identifying critical LSM parameters influencing the results of NWP models, focusing in particular on the simulation of thermally-driven circulations over complex terrain. While previous sensitivity analyses employed offline LSM simulations to evaluate the sensitivity to surface parameters, this study adopts an online coupled approach, utilizing the Noah-MP LSM within the Weather Research and Forecasting (WRF) model. To overcome computational constraints, a novel tool, Machine Learning-based Automated Multi-method Parameter Sensitivity and Importance analysis Tool (ML-AMPSIT), is developed and tested. This tool allows users to explore the sensitivity of the results to model parameters using supervised machine learning regression algorithms, including Random Forest, CART, XGBoost, SVM, LASSO, Gaussian Process Regression, and Bayesian Ridge Regression. These algorithms serve as fast surrogate models, greatly accelerating sensitivity analyses while maintaining a high level of accuracy. The versatility and effectiveness of ML-AMPSIT enable the fast implementation of advanced sensitivity methods, such as the Sobol method, overcoming the computational limitations encountered in expensive models like WRF. The suitability of this tool to assess model’s sensitivity to the variation of specific parameters is first tested in an idealized sea breeze case study where six surface parameters are varied. Then, the analysis focuses on the evaluation of the sensitivity to surface parameters in the simulation of thermally-driven circulations in a mountain valley. Specifically, an idealized three-dimensional topography consisting of a valley-plain system is adopted, analyzing a complete diurnal cycle of valley and slope winds. The analysis focuses on all the key surface parameters governing the interactions between NoahMP and WRF. The proposed approach, novel in the context of LSM-NWP model coupling, draws from established applications of machine learning in various Earth science disciplines, underscoring its potential to improve the estimation of parameter sensitivities in NWP models.

[ES] El desarrollo del medio digital ha transformado nuestra forma de consumo en las últimas décadas. La invención de Internet, su democratización, la aparición de múltiples dispositivos de acceso y las redes sociales, la tecnificación de los objetos y la llegada de la inteligencia artificial han tenido un impacto significativo en la sociedad, así como entidades esenciales como las empresas y sus marcas. La integración total de la digitalización en las marcas es una realidad, y cada vez se opta más por este medio como un espacio prioritario para aportar valor al público a través de sus productos o servicios. Esta investigación se centra en profundizar en el significado de la marca digital y en sus características esenciales. Para ello, se realizará un recorrido histórico de la evolución de los signos identitarios con relación a la tecnología, lo que permitirá tener un enfoque global en su adaptación a cada uno de los avances digitales. Además, se analizarán los múltiples significados de marca y se revisará y recogerá la metodología específica para su creación: el branding. Con el fin de entender las particularidades y ventajas de los marcos de trabajo aplicados en el sector digital y del desarrollo de software, se estudiarán metodologías iterativas basadas en sistemas ágiles como el Design Thinking, el Diseño Centrado en Usuario o el Atomic Design, entre otros. Finalmente, a partir del estudio realizado, se generará una metodología híbrida para crear marcas digitales capaces de adaptarse mejor a los cambios de contexto del medio. Para ello, se hará uso de procesos, herramientas y plataformas complementarias empleadas en ámbitos tecnológicos y se diseñará un proceso de revisión constante con el fin de asegurar la calidad y el buen funcionamiento de las marcas en todo momento. / [CA] El desenvolupament dels mitjans digitals han transformat la nostra forma de consum en les últimes dècades. La invenció d'Internet, la seua democratització, l'aparició de múltiples dispositius d'accés i les xarxes socials, la tecnificació dels objectius i l'arribada de la intel·ligència artificial han tingut un impacte significatiu en la societat i en les entitats essencials com les empreses i les seues marques. La integració total de la digitalització en les marques és una realitat, i cada vegada s'opta més per aquest mitjà com un espai prioritari per a aportar valor al públic. Aquesta investigació es centra en aprofundir en el significat de la marca digital i en les seues característiques essencials. Per a això, es realitzarà un recorregut històric de l'evolució dels signes identitaris en relació amb la tecnologia, la qual cosa permetrà tindre un enfocament global de la seua adaptació a cadascun dels sorgiments digitals. A més a més, s'analitzaran els múltiples significats de marca i es revisarà i recollirà la metodologia específica per a la seua creació: el branding. Amb la finalitat d'entendre les particularitats i avantatges dels marcs de treball aplicats al sector digital i del desenvolupament del software, s'estudiaran metodologies iteratives basades en sistemes àgils com el Design Thinking, el Disseny Centrat en l'Usuari, l'Atomic Design, entre d'altres. Finalment, a partir de l'estudi realitzat, es generarà una metodologia híbrida per a crear marques digitals capaces d'adaptar-se millor als canvis de context del mitjà. Per a això, es farà ús dels processos, eines i plataformes complementàries emprades en àmbits tecnològics i es dissenyarà un procés de revisió constant amb la finalitat d'assegurar la qualitat i el bon funcionament de les marques en tot moment. / [EN] The advancement of digital media has significantly impacted how we consume information in recent years. With the Internet and social networks becoming more accessible, coupled with the emergence of multiple access devices and the application of artificial intelligence, society and essential entities such as companies and their brands have been significantly affected. Digitalisation has become an integral part of branding, and companies now prioritize using digital media to provide value to their customers. This research explores the meaning of digital branding and its fundamental characteristics. It will provide a historical overview of how identity signs have evolved with technological advancements, offering a comprehensive approach to their adaptation in the digital age. To fully understand the advantages and nuances of digital and software development frameworks, this study will delve into iterative methodologies based on agile systems, such as Design Thinking, User-Centered Design, and Atomic Design. Ultimately, the study will generate a hybrid methodology for creating digital brands that can adapt better to environmental changes. For this purpose, other complementary processes, such as tools and platforms used in technological fields, will be used. A constant review process will also be present to ensure the quality and proper functioning of the brands at all times. / Canavese Arbona, A. (2023). Diseño de identidades digitales: metodología iterativa para la creación y desarrollo de marcas [Tesis doctoral]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/196737

This dissertation uses spatial filtering to develop a large eddy simulation reduced order model (LES-ROM) framework for fluid flows. Proper orthogonal decomposition is utilized to extract the dominant spatial structures of the system. Within the general LES-ROM framework, two approaches are proposed to address the celebrated ROM closure problem. No phenomenological arguments (e.g., of eddy viscosity type) are used to develop these new ROM closure models. The first novel model is the approximate deconvolution ROM (AD-ROM), which uses methods from image processing and inverse problems to solve the ROM closure problem. The AD-ROM is investigated in the numerical simulation of a 3D flow past a circular cylinder at a Reynolds number $Re=1000$. The AD-ROM generates accurate results without any numerical dissipation mechanism. It also decreases the CPU time of the standard ROM by orders of magnitude. The second new model is the calibrated-filtered ROM (CF-ROM), which is a data-driven ROM. The available full order model results are used offline in an optimization problem to calibrate the ROM subfilter-scale stress tensor. The resulting CF-ROM is tested numerically in the simulation of the 1D Burgers equation with a small diffusion parameter. The numerical results show that the CF-ROM is more efficient than and as accurate as state-of-the-art ROM closure models. / Ph. D. / Numerical simulation of complex fluid flows is often challenging in many realistic engineering, scientific, and medical applications. Indeed, an accurate numerical approximation of such flows generally requires millions and even billions of degrees of freedom. Furthermore, some design and control applications involve repeated numerical simulations for different parameter values. Reduced order models (ROMs) are an efficient approach to the numerical simulation of fluid flows, since they can reduce the computational time of a brute force computational approach by orders of magnitude while preserving key features of the flow. Our main contribution to the field is the use of spatial filtering to develop better ROMs. To construct the new spatially filtered ROMs, we use ideas from image processing and inverse problems, as well as data-driven algorithms. The new ROMs are more accurate than standard ROMs in the numerical simulation of challenging three-dimensional flows past a circular cylinder.

<p dir="ltr">This thesis focusses on developing novel data driven oral drug formulation analysis methods by employing technologies such as Fourier transform analysis and generative adversarial learning. Data driven measurements have been addressing challenges in advanced manufacturing and analysis for pharmaceutical development for the last two decade. Data science combined with analytical chemistry holds the future to solving key problems in the next wave of industrial research and development. Data acquisition is expensive in the realm of pharmaceutical development, and how to leverage the capability of data science to extract information in data deprived circumstances is a key aspect for improving such data driven measurements. Among multiple measurement techniques, chemical imaging is an informative tool for analyzing oral drug formulations. However, chemical imaging can often fall into data deprived situations, where data could be limited from the time-consuming sample preparation or related chemical synthesis. An integrated imaging approach, which folds data science techniques into chemical measurements, could lead to a future of informative and cost-effective data driven measurements. In this thesis, the development of data driven chemical imaging techniques for the analysis of oral drug formulations via Fourier transformation and generative adversarial learning are elaborated. Chapter 1 begins with a brief introduction of current techniques commonly implemented within the pharmaceutical industry, their limitations, and how the limitations are being addressed. Chapter 2 discusses how Fourier transform fluorescence recovery after photobleaching (FT-FRAP) technique can be used for monitoring the phase separated drug-polymer aggregation. Chapter 3 follows the innovation presented in Chapter 1 and illustrates how analysis can be improved by incorporating diffractive optical elements in the patterned illumination. While previous chapters discuss dynamic analysis aspects of drug product formulation, Chapter 4 elaborates on the innovation in composition analysis of oral drug products via use of novel generative adversarial learning methods for linear analyses.</p>

Vehicle monitoring is an important prequisite for predictive maintenance applications. Virtual sensors can be deployed to establish relationships between fatigue related quantities of interest and readily available measurement data, which reduces the costs of monitoring for vehicle fleets. This work describes a data-driven virtual sensing approach using the scattering transform and principal component analysis. These data transformations are used to obtain a reduced representation of acceleration data, which is suitable for the identification of fatigue critical events during vehicle operation. Results of a previous study using an eBike demonstrator are summarized and the methodology is applied to experimental data of a sensor equipped light rail vehicle. In both applications, fictitious fatigue damage contributions are estimated accurately and physical interpretations of the reduced representation are found.

This dissertation presents a data-driven computational epidemic framework to simulate disease epidemics at global mass gatherings. The annual Muslim pilgrimage to Makkah, Saudi Arabia is used to demonstrate the simulation and analysis of various disease transmission scenarios throughout the different stages of the event from the arrival to the departure of international participants. The proposed agent-based epidemic model efficiently captures the demographic, spatial, and temporal heterogeneity at each stage of the global event of Hajj. Experimental results indicate the substantial impact of the demographic and mobility patterns of the heterogeneous population of pilgrims on the progression of the disease spread in the different stages of Hajj. In addition, these simulations suggest that the differences in the spatial and temporal settings in each stage can significantly affect the dynamic of the disease. Finally, the epidemic simulations conducted at the different stages in this dissertation illustrate the impact of the differences between the duration of each stage in the event and the length of the infectious and latent periods. This research contributes to a better understanding of epidemic modeling in the context of global mass gatherings to predict the risk of disease pandemics caused by associated international travel. The computational modeling and disease spread simulations in global mass gatherings provide public health authorities with powerful tools to assess the implication of these events at a different scale and to evaluate the efficacy of control strategies to reduce their potential impacts.

Die Physik des Top-Quarks ist eine Schlüsselkomponente im Forschungsprogramm des ATLAS-Experiments am CERN. In dieser Arbeit werden Untersuchungen zur Leistungfähigkeit von Jet-Triggern für Top-Quark-Ereignisse präsentiert und zwei datenbasierte Methoden zur Abschätzung der Multijet-Triggereffizienz und des W+Jets-Untergrundes in Top-Quark-Ereignissen in ATLAS eingeführt. In einer tag-and-probe Methode, basierend auf einer einfachen und allgemeinen Ereignisselektion und einem hochenergetischen Lepton als Tag, wird die Möglichkeit zur Bestimmung der Multijet-Triggereffizienz aus Daten heraus evaluiert, und es wird gezeigt, dass die Methode in der Lage ist, die Effizienz ohne signifikante Verfälschung durch die Tag-Selektion zu bestimmen. In der zweiten datenbasierten Analyse wird eine neue Methode zur Abschätzung des W+Jets-Untergrundes in ATLAS eingeführt. Durch die Definition von signal- und untergrunddominierten Bereichen in Jet-Muliplizität und Pseudorapidität des Leptons wird der Anteil der W+Jets-Ereignisse aus der untergrunddominierten in die signaldominierte Region extrapoliert. Es wird gezeigt, dass die Methode, mit einer integrierten Luminosität von 100 pb^−1 bei sqrt(s) = 10 TeV, in der Lage ist den Untergrundbeitrag als Funktion der Jet-Muliplizität mit etwa 25% Genauigkeit im Großteil der signaldominierten Region zu bestimmen. Diese Arbeit umfaßt zudem eine Studie zum thermischen Verhalten und der erwarteten thermischen Leistung des Pixel-Detektors in ATLAS. Alle Messungen, durchgeführt während der Inbetriebnahme des Systems in 2008/09, zeigen Ergebnisse innerhalb der Spezifikationen beziehungweise deuten auf deren Einhaltung auch nach mehreren Betriebsjahren unter LHC-Bedingungen hin. / The physics of the top quark is one of the key components in the physics programme of the ATLAS experiment at the Large Hadron Collider at CERN. In this thesis, general studies of the jet trigger performance for top quark events using fully simulated Monte Carlo samples are presented and two data-driven techniques to estimate the multi-jet trigger efficiency and the W+Jets background in top pair events are introduced to the ATLAS experiment. In a tag-and-probe based method, using a simple and common event selection and a high transverse momentum lepton as tag object, the possibility to estimate the multijet trigger efficiency from data in ATLAS is investigated and it is shown that the method is capable of estimating the efficiency without introducing any significant bias by the given tag selection. In the second data-driven analysis a new method to estimate the W+Jets background in a top-pair event selection is introduced to ATLAS. By defining signal and background dominated regions by means of the jet multiplicity and the pseudo-rapidity distribution of the lepton in the event, the W+Jets contribution is extrapolated from the background dominated into the signal dominated region. The method is found to estimate the given background contribution as a function of the jet multiplicity with an accuracy of about 25% for most of the top dominated region with an integrated luminosity of above 100 pb^−1 at sqrt(s) = 10 TeV. This thesis also covers a study summarising the thermal behaviour and expected performance of the Pixel Detector of ATLAS. All measurements performed during the commissioning phase of 2008/09 yield results within the specification of the system and the performance is expected to stay within those even after several years of running under LHC conditions.

Dans le cadre du projet « InterPhonologie du Français Contemporain » (IPFC) (Racine, Detey et Kawaguchi 2012) ce travail se propose d’examiner les stratégies d’acquisition de la liaison en L2. Si des modèles développementaux ont été proposés pour l’acquisition de la liaison en L1, aucune hypothèse pour rendre compte de l’apprentissage en L2 n’a reçu à ce jour un appui empirique convaincant (Wauquier 2009). C’est dans ce contexte que s’inscrit la présente étude longitudinale menée auprès d’étudiants italophones de français langue étrangère (FLE) (niveau A2-B1) dont les résultats ont été enrichis par une comparaison avec les autres travaux multilingues issus du projet international IPFC. Cette approche a permis de mettre en évidence la présence de tendances et d’erreurs communes qui suggèrent que, malgré le recours à des stratégies lexicales, les apprenants sont en mesure de développer des généralisations phonologiques de la liaison. De plus, compte-tenu des difficultés que l’hétérogénéité de la liaison pose dans l’enseignement et l'acquisition en français L2 (Racine et Detey sous presse), les apprenants montrent des faiblesses tant au niveau de la production de la liaison qu’au niveau des connaissances épilinguistiques (Gombert 1990) de la variation de la liaison. Ces résultats nous ont amené à exploiter les données extraites de l’analyse fréquentielle du corpus des natifs « Phonologie du Français Contemporain » (PFC) (Durand, Laks et Lyche 2009) afin de proposer des ressources pédagogiques dont nous espérons qu’elles seront en mesure, dans le cadre d’une démarche data-driven learning, de contribuer au renouvèlement de l′enseignement de la liaison en français langue étrangère. / As part of the international project InterPhonologie du Français Contemporain (IPFC) (Detey Kawaguchi and 2008; Racine, Detey and Kawaguchi 2012), this study aims to examine acquisition strategies of liaison in French as a foreign language. While developmental models have been proposed for the acquisition of liaison in L1, to date, no hypothesis accounting for L2 liaison learning has received convincing empirical support (Wauquier 2009). It is in this context that the present longitudinal study of Italian-speaking students of French as a foreign language (FFL) (level A2-B1) can be situated. The results have been enriched by a comparison with other multilingual studies (several source languages) from the international project IPFC. This approach shows the presence of common trends and errors in different populations of learners that suggest that, despite the use of lexical strategies, learners are able to develop phonological generalisations of liaison. Taking into account the difficulties that the heterogeneity of liaison presents for second language teaching and acquisition (Racine and Detey in press), learners show weaknesses in both production of liaison and in epilinguistic knowledge (Gombert 1990) of the variation of liaison.These results led us to use data from a frequency analysis of a corpus of native speakers Phonologie du Français Contemporain (PFC) (Durand, Laks and Lyche 2009) to provide learning resources, based on a data-driven learning approach, to contribute to the renewal of the teaching of liaison in French as a foreign language. / Nata nell’ambito del progetto InterPhonologie du Français Contemporain (IPFC) (Racine, Detey et Kawaguchi 2012), questa tesi propone un’analisi delle strategie di acquisizione della liaison in L2. Sebbene dei modelli di acquisizione della liaison in L1 siano già stati proposti, nessuna ipotesi è stata confermata da studi empirici (Wauquier 2009).È in questo contesto che si situa il nostro studio longitudinale realizzato su degli studenti universitari italofoni di francese lingua straniera (FLS) (livello A2-B1); un’osservazione i cui risultati sono stati confrontati con altre ricerche multilingui facenti parte del progetto internazionale IPFC. Questo approccio ha permesso di far emergere la presenza di tendenze e errori comuni alle diverse popolazioni di apprendenti; una similitudine che suggerisce come gli apprendenti, malgrado ricorrano a diverse strategie lessicali, sono in grado di sviluppare delle generalizzazioni fonologiche della liaison. Inoltre, date le difficoltà che la variabilità della liaison pone tanto nella didattica quanto nell’acquisizione, gli apprendenti sembrano possedere una scarsa competenza della variazione della liaison non solo sul piano della produzione ma anche su quello delle conoscenze epilinguistiche (Gombert 1990).Questi risultati ci hanno spinto quindi ad utilizzare i dati dell’analisi della frequenza della liaison condotta sul corpus di nativi Phonologie du Français Contemporain (PFC) (Durand, Laks et Lyche 2009) al fine di proporre delle risorse pedagogiche per l’acquisizione della liaison. Questi strumenti potrebbero contribuire al rinnovamento dell’insegnamento della liaison nella didattica del francese lingua straniera.

[pt] Otimização de portfólio tradicionalmente assume ter conhecimento da distribuição de probabilidade dos retornos ou pelo menos algum dos seus momentos. No entanto, é sabido que a distribuição de probabilidade dos retornos muda com frequência ao longo do tempo, tornando difícil a utilização prática de modelos puramente estatísticos, que confiam indubitavelmente em uma distribuição estimada. Em contrapartida, otimização robusta considera um completo desconhecimento da distribuição dos retornos, e por isto, buscam uma solução ótima para todas as realizações possíveis dentro de um conjunto de incerteza dos retornos. Mais recentemente na literatura, técnicas de distributionally robust optimization permitem lidar com a ambiguidade com relação à distribuição dos retornos. No entanto essas técnicas dependem da construção do conjunto de ambiguidade, ou seja, distribuições de probabilidade a serem consideradas. Neste trabalho, propomos a construção de conjuntos de ambiguidade poliédricos baseado somente em uma amostra de retornos. Nestes conjuntos, as relações entre variáveis são determinadas pelos dados de maneira não paramétrica, sendo assim livre de possíveis erros de especificação de um modelo estocástico. Propomos um algoritmo para construção do conjunto e, dado o conjunto, uma reformulação computacionalmente tratável do problema de otimização de portfólio. Experimentos numéricos mostram que uma melhor performance do modelo em comparação com benchmarks selecionados. / [en] Portfolio optimization traditionally assumes knowledge of the probability distribution of returns or at least some of its moments. However is well known that the probability distribution of returns changes over time, making difficult the use of purely statistic models which undoubtedly rely on an estimated distribution. On the other hand robust optimization consider a total lack of knowledge about the distribution of returns and therefore it seeks an optimal solution for all the possible realizations wuthin a set of uncertainties of the returns. More recently the literature shows that distributionally robust optimization techniques allow us to deal with ambiguity regarding the distribution of returns. However these methods depend on the construction of the set of ambiguity, that is, all distribution of probability to be considered. This work proposes the construction of polyhedral ambiguity sets based only on a sample of returns. In those sets, the relations between variables are determined by the data in a non-parametric way, being thus free of possible specification errors of a stochastic model. We propose an algorithm for constructing the ambiguity set, and then a computationally treatable reformulation of the portfolio optimization problem. Numerical experiments show that a better performance of the model compared to selected benchmarks.