HAL Portal UO (Université d'Orléans)
Not a member yet
48841 research outputs found
Sort by
Amélioration des stratégies de portefeuille crédit : perspectives sur les opérations sur titres, les algorithmes d'apprentissage du classement et la couverture des taux d’intérêt
No full textThe evolution of credit markets, driven by regulatory reforms, trading electronification, and the increasing availability of data, has reshaped the landscape for asset managers. This thesis contributes to the advancement of systematic credit strategies by investigating credit risk dynamics, machine learning model design, and interest rate hedging techniques.The first contribution examines the impact of corporate action announcements—such as stock buybacks, dividends, and mergers & acquisitions—on credit default swap spreads. Quantifying abnormal spread movements around these events allows for the disentangling of wealth transfer from signaling effects, providing insights into the mechanisms driving credit market reactions.The second contribution focuses on improving asset ranking models for long-short portfolio construction. Traditional learning-to-rank algorithms favor top-ranked assets while neglecting underperformers, creating challenges for market-neutral strategies. This thesis proposes refined loss functions that enhance ranking symmetry across the full asset distribution, thereby improving portfolio balance and performance.The third contribution introduces a multi-task neural network framework that unifies asset ranking and portfolio weighting within a single optimization model. In contrast to conventional two-step approaches, this framework simultaneously optimizes both ranking and allocation, resulting in more efficient investment decisions.Finally, this thesis explores the relationship between equity and corporate bond markets by decomposing their correlation into credit risk and interest rate sensitivity components. Since these drivers vary across credit qualities and market conditions, this thesis proposes two adaptive interest rate hedging strategies. These strategies outperform traditional duration-convexity hedging by dynamically adjusting hedge ratios in response to evolving market regimes.These contributions collectively advance systematic credit investing by integrating financial expertise with advanced quantitative techniques. The findings offer asset managers practical methodologies for improving credit trading strategies in an increasingly data-driven and complex financial environment.L’évolution du marché du crédit, portée par les réformes réglementaires, l’électronification des passages d'ordre et l’augmentation des données disponibles, a profondément modifié les pratiques des gestionnaires d'actifs. Cette thèse contribue à l’amélioration des stratégies de crédit systématiques en explorant la dynamique du risque de crédit, la conception de modèles d’apprentissage automatique et les techniques de couverture des taux d’intérêt.La première contribution analyse l’impact des annonces d’opérations sur titres—telles que les rachats d’actions, les dividendes et les fusions-acquisitions—sur les spreads des credit default swaps. En quantifiant les mouvements anormaux des spreads autour de ces événements, il a été possible de dissocier l'effet "wealth transfer" de l'effet "signaling", offrant ainsi une meilleure compréhension des mécanismes sous-jacents aux réactions du marché du crédit.La deuxième contribution améliore les modèles de classement d'actifs pour la construction de portefeuilles long-short. Les algorithmes traditionnels d’apprentissage du classement favorisent les actifs les mieux classés au détriment des moins performants, rendant les stratégies neutres au marché moins efficaces. Cette thèse propose des fonctions de perte ajustées afin de renforcer la symétrie du classement sur l’ensemble de la distribution des actifs, améliorant ainsi l’équilibre et la performance des portefeuilles.La troisième contribution propose un modèle de réseau neuronal multitâche unifiant le classement des actifs et la pondération des portefeuilles dans un modèle d’optimisation unique. Contrairement aux méthodes conventionnelles en deux étapes, cette approche produit une construction de portefeuille plus efficiente.Enfin, cette thèse examine la relation entre les marchés actions et des obligations d’entreprise en décomposant leur corrélation en deux composantes : le risque de crédit et la sensibilité aux taux d’intérêt. Puisque ces facteurs varient selon les qualités de crédit et les conditions de marché, deux stratégies adaptatives de couverture des taux d’intérêt sont proposées. Ces stratégies surpassent la couverture traditionnelle "duration-convexity" en ajustant dynamiquement les ratios de couverture selon les régimes de marché.Les contributions de cette thèse offrent aux gestionnaires d’actifs de nouvelles méthodologies afin d'optimiser les stratégies de gestion sur le marché du crédit, dans un environnement financier toujours plus complexe et axé sur les données
Degenerate Elliptic PDEs on a Network with Kirchhoff Conditions
No full textIn this article, we are interested in semilinear, possibly degenerate elliptic equations posed on a general network, with nonlinear Kirchhoff-type conditions for its interior vertices and Dirichlet boundary conditions for the boundary ones. The novelty here is the generality of the equations posed on each edge that is incident to a particular vertex, ranging from first-order equations to uniformly elliptic ones. Our main result is a strong comparison principle, i.e., a comparison result between discontinuous viscosity sub and supersolutions of such problems, from which we conclude the existence and uniqueness of a continuous viscosity by Perron's method. Further extensions are also discussed
Attentes des médecins généralistes vis-à-vis des kinésithérapeutes libéraux : une étude Delphi sur la communication interprofessionnelle
No full textInternational audienceIntroductionInterprofessional collaboration is essential to the quality, continuity and efficiency of care. Yet communication between general practitioners (GPs) and physiotherapists remains limited. Existing tools, such as the BDK form or the SAED model, are little used, as they are perceived as too technical or unsuited to daily medical practice. The aim of this study was to identify the information that GPs consider relevant to receive from KLs, in order to build a transmission model better adapted to their expectations.MethodsA modified Delphi study was conducted in a pragmatic paradigm. Twelve GPs took part in the first round, nine in the next two. Three waves of questionnaires were distributed, with consensus thresholds set at 80% for the first two rounds and 70% for the third. Anonymity, iterability and controlled feedback were guaranteed via the Lime Survey© platform. Items were constructed from scientific, regulatory and clinical sources, then validated by pre-testing.ResultsTwenty-six items reached consensus. GPs preferred targeted, non-systematic transmission, to be sent at the end of care or in the event of an adverse event. The preferred format was a one-page summary sheet sent by secure messaging system. The BDK form, considered too dense, was rejected. Information deemed relevant included red flags, clinical evolution, pain (scales, location, modulating factors), functional limitations, patient motivation and compliance, patient goals and, if necessary, those of the KL, as well as frequency of sessions and influential treatments.DiscussionThe results confirm that GPs expect communication to be synthetic and directly useful for medical decision-making, rather than a detailed account of the physiotherapist's reasoning process. This position reflects an evolution in the interprofessional relationship, valuing the KL's expertise in a collaborative framework. The prescription is thus perceived less as an order than as a coordination tool. The study also highlights the fact that the form of brief transmission integrated with digital tools determines its acceptability. GPs prefer secure channels that can be integrated into their business software. Paper, telephone or DMP-based transmissions are rarely desired.ConclusionThis study highlights a clear consensus among GPs in favor of synthetic, structured interprofessional communication. It proposes a co-constructed model aligned with their expectations, to be integrated into KLs’ daily practice. This dynamic opens up concrete prospects for strengthening collaboration, adapting digital tools, and providing input for initial and continuing training in interprofessionality.IntroductionLa collaboration interprofessionnelle est essentielle à la qualité, la continuité et l’efficience des soins. Pourtant, la communication entre médecins généralistes (MG) et kinésithérapeutes libéraux (KL) reste limitée. Les outils existants, tels que la fiche bilan diagnostique kinésithérapique (BDK) ou le modèle situation-antécédents-évaluation-demande (SAED), sont peu utilisés car perçus comme trop techniques ou inadaptés à la pratique médicale quotidienne. Cette étude visait à identifier les informations que les MG jugent pertinentes à recevoir de la part des KL afin de construire un modèle de transmission mieux adapté à leurs attentes.MéthodesUne étude Delphi modifiée a été menée dans un paradigme pragmatique. Douze MG ont participé au premier tour, neuf aux deux suivants. Trois vagues de questionnaires ont été diffusées, avec des seuils de consensus fixés à 80 % pour les deux premières rondes et à 70 % pour la troisième. L’anonymat, l’itérativité et un retour contrôlé d’information ont été garantis via la plateforme Lime Survey©. Les items ont été construits à partir de sources scientifiques, réglementaires et cliniques, puis validés par un pré-test.RésultatsVingt-six items ont atteint le consensus. Les MG privilégiaient une transmission ciblée, non systématique, à adresser en fin de prise en soins ou en cas d’événement indésirable. Le format préféré était une fiche synthétique d’une page transmise par messagerie sécurisée. La fiche BDK, jugée trop dense, était rejetée. Les informations jugées pertinentes incluaient les drapeaux rouges, l’évolution clinique, la douleur (échelles, localisation, facteurs modulateurs), les limitations fonctionnelles, la motivation et l’observance du patient, les objectifs du patient et, si nécessaire, ceux du KL, ainsi que la fréquence des séances et les traitements influents.DiscussionLes résultats confirment que les MG attendent une communication synthétique et directement utile à la décision médicale, et non un compte rendu détaillé du processus de raisonnement du kinésithérapeute. Cette position reflète une évolution dans la relation interprofessionnelle qui valorise l’expertise du KL dans un cadre collaboratif. Ainsi, la prescription est perçue moins comme un ordre que comme un outil de coordination. L’étude souligne également que la forme de la transmission brève, intégrée aux outils numériques, conditionne son acceptabilité. Les MG privilégient les canaux sécurisés intégrables à leur logiciel métier. Les transmissions papier, téléphoniques ou via le DMP sont peu souhaitées.ConclusionCette étude montre un consensus net des MG en faveur d’une communication interprofessionnelle synthétique et structurée. Elle propose un modèle co-construit aligné sur leurs attentes, à intégrer dans la pratique quotidienne des KL. Cette dynamique ouvre des perspectives concrètes pour renforcer la collaboration, adapter les outils numériques, et alimenter les dispositifs de formation initiale et continue à l’interprofessionnalité
Physico-mechanical characterization of 3D-printed PLGA for patient-specific resorbable implants in craniofacial surgery
No full textInternational audiencePatient-specific implant (PSI) has optimized the management for a wide range of complex craniofacial deformity over the past years by increasing the accuracy of surgical procedures and lowering the operating time. In hypertelorism (HTO) surgery particularly, the orbital bone repositioning is nowadays guided by patient-specific bone fixation plates that are usually made from non-resorbable alloplastic material (e.g., titanium). Developing resorbable personalized plates could be a relevant alternative to overcome the well-known drawbacks of titanium plates such as infection, exposure or even the lack of bone growth which is detrimental in pediatric patients. This study investigated the mechanical and structural characteristics of poly(lactic-co-glycolic acid) (PLGA) PSI as resorbable materials for HTO surgery. We assessed the feasibility of printing PLGA PSI by Fused Deposition Modeling additive manufacturing (FDM). The geometrical and the mechanical properties of the 3D-printed device were compared with standard resorbable plates and analyzed after sterilization process (i.e., hydrogen peroxide gas plasma). The Young’s modulus was greater than the standard resorbable plates while a decrease of 36% (p = 0.004) after the sterilization was observed. The sterilization also induced a plate deformation with an increase of 0.27 mm in Z-axis and a decrease of 0.8 mm in Y-axis due to annealing effect. Compared to the design, the PLGA PSI were successfully 3D-printed with a maximum deviation of 0.1 mm, making our custom-made plate promising for personalized craniofacial applications. Further investigations on the sterilization process must be considered in view of its mechanical and structural impact on resorbable PSI
Resilient math inspired EDA optimized fuzzy adaptive exponent controller for LFC improvement of an EV integrated microgrid: This study aims to stabilize the frequency of an electric vehicle-integrated AC microgrid in different electrical uncertainties. The recommended microgrid is built by incorporating different distributed generation (DG) oriented power plants. The DG system includes a wind power plant, solar PV plant, diesel generator, fuel cell and geothermal plant. The microgrid frequency goes on oscillating under the action of few uncertainties like dynamics in applied load, fluctuation in wind power and variability in solar power intensity. Further, the charging of electric vehicles extremely disturbs the grid frequency and causes frequency instability issues in the microgrid. This proposed study has anticipated a Fuzzy adaptive exponent PID (Fuzzy PI-DÆ) controller to obtain stability in microgrid frequency under different disturbances. Further, the microgrid is associated with different energy-storing devices for improving overall power quality of the system. The Fuzzy PI-DÆ parameters are selected in optimum by incorporating an advanced Math inspired-Exponential distribution algorithm (Mi-EDA) in different operations. The potential of the optimal Fuzzy PI-DÆ controller is compared with fractional ordered fuzzy PID (FO-FPID), Fuzzy PID and PID controllers in concern to the microgrid’s frequency stabilization. The research findings conclude that, the anticipated Fuzzy PI-DÆ approach promptly advances the settling time of frequency by 72.72% and 136.32% and 345.46% to that of FO-FPID, Fuzzy PID and PID controllers respectively. The optimal property of the recommended Mi-EDA technique is compared with the typical sine cosine algorithm (SCA), GA and PSO techniques for validating the potential of the technique.
No full textInternational audienceAbstract This study aims to stabilize the frequency of an electric vehicle-integrated AC microgrid in different electrical uncertainties. The recommended microgrid is built by incorporating different distributed generation (DG) oriented power plants. The DG system includes a wind power plant, solar PV plant, diesel generator, fuel cell and geothermal plant. The microgrid frequency goes on oscillating under the action of few uncertainties like dynamics in applied load, fluctuation in wind power and variability in solar power intensity. Further, the charging of electric vehicles extremely disturbs the grid frequency and causes frequency instability issues in the microgrid. This proposed study has anticipated a Fuzzy adaptive exponent PID (Fuzzy PI-D Æ ) controller to obtain stability in microgrid frequency under different disturbances. Further, the microgrid is associated with different energy-storing devices for improving overall power quality of the system. The Fuzzy PI-D Æ parameters are selected in optimum by incorporating an advanced Math inspired-Exponential distribution algorithm (Mi-EDA) in different operations. The potential of the optimal Fuzzy PI-D Æ controller is compared with fractional ordered fuzzy PID (FO-FPID), Fuzzy PID and PID controllers in concern to the microgrid’s frequency stabilization. The research findings conclude that, the anticipated Fuzzy PI-D Æ approach promptly advances the settling time of frequency by 72.72% and 136.32% and 345.46% to that of FO-FPID, Fuzzy PID and PID controllers respectively. The optimal property of the recommended Mi-EDA technique is compared with the typical sine cosine algorithm (SCA), GA and PSO techniques for validating the potential of the technique
Quand la com’ s’emballe : le feu de paille du « starter pack »
No full textInternational audienceAdoptés par de nombreuses marques, les visuels de figurines sous starter pack ont rapidement envahi les réseaux sociaux. Derrière cette stratégie virale en apparence efficace se cache une communication mimétique, simplificatrice qui s’est rapidement trouvée en situation d’essoufflement. Une tendance qui en dit long sur les dérives de la communication à l’ère de l’instantanéité
Une recherche translationnelle: De la paillasse aux patients
No full textInternational audienceCette journée a permis aux patients atteints de sclérose en plaques et à leur famille de comprendre comment se déroule un projet de recherche et quelles perspectives émergent des travaux scientifiques. Les chercheurs ont pu échanger avec des personnes vivant la maladie au quotidien et comprendre leurs difficultés et attentes.Les visiteurs ont participé à deux conférences:-Une conférence du Dr Maud Allix-Guyot, neurologue au CHU d’Orléans-Une présentation de l’équipe « Neurobiologie des récepteurs et des RNAi pour des innovations thérapeutiques » et de ses recherches sur la sclérose en plaques par le Dr Séverine Morisset-Lopez, chercheuse CNRS -Quatre ateliers ont été proposés:o« Comment les chercheurs cultivent des cellules pour étudier la sclérose en plaques ? »o« Les grandes étapes pour identifier de nouveaux candidats médicaments » o« parcours d’un échantillon de sang : du patient à la paillasse pour de nouvelles découverteso« Comment visualiser les cellules de notre cerveau ?
Removing motion artifacts from mechanomyographic signals: an innovative filtering method applied to human movement analysis
No full textInternational audienceMechanomyography (MMG) is a promising tool for measuring muscle activity in the field but its sensitivity to motion artifacts limits its application. In this study, we proposed an adaptative filtering method for MMG accelerometers based on the complete ensemble empirical mode decomposition, with adaptative noise and spectral fuzzy entropy, to isolate motions artefacts from the MMG signal in dynamic conditions. We compared our method with the traditional band-pass filtering technique, demonstrating better results concerning motion recomposition for deltoid and erector spinae muscles (R² = 0.907 and 0.842). Thus, this innovative method allows the filtering of motion artifacts dynamically in the 5-20 Hz bandwidth, which is not achievable with traditional method. However, the interpretation of accelerometric MMG signals from the trunk and lower-limb muscles during walking or running should be approached with great caution as impact-related accelerations are still present, though their exact quantity still needs to be quantified
French Lexicography
No full textInternational audienceThis article explores the history and evolution of French lexicography, from its early beginnings in the medieval glossaries to the foundational works of the 17th century, such as the first monolingual French dictionaries.It traces the creation of pivotal lexicographic models, including those by Pierre Richelet, Antoine Furetière, and the French Academy, which laid the groundwork for modern dictionaries. The article also discusses the rise of digital lexicography in the 21st century, emphasizing its continuous development and adaptation, including the ongoing work of the ninth edition of theDictionnaire de l’Académie française
Islands for black holes in a hybrid quantum state
No full textInternational audienceFollowing our previous work on hybrid quantum states in the RST model, we study its most interesting solution representing a completely regular spacetime with the structure of causal diamond, containing an apparent horizon and radiation at infinity. Adapting recent computations of radiation entropy in terms of the entropy of entanglement, we find that this entropy follows a Page curve. This confirms our previous result [1], which was obtained by directly calculating the thermodynamic entropy of radiation at infinity. We also investigate the presence of a possible island in these systems, and find that it does not seem to play a role in contributing to the generalized black hole entropy