Rozprawy doktorskie na temat „Modèles prédictif”
Utwórz poprawne odniesienie w stylach APA, MLA, Chicago, Harvard i wielu innych
Sprawdź 50 najlepszych rozpraw doktorskich naukowych na temat „Modèles prédictif”.
Przycisk „Dodaj do bibliografii” jest dostępny obok każdej pracy w bibliografii. Użyj go – a my automatycznie utworzymy odniesienie bibliograficzne do wybranej pracy w stylu cytowania, którego potrzebujesz: APA, MLA, Harvard, Chicago, Vancouver itp.
Możesz również pobrać pełny tekst publikacji naukowej w formacie „.pdf” i przeczytać adnotację do pracy online, jeśli odpowiednie parametry są dostępne w metadanych.
Przeglądaj rozprawy doktorskie z różnych dziedzin i twórz odpowiednie bibliografie.
Benson, Marie Anne. "Pouvoir prédictif des données d'enquête sur la confiance". Master's thesis, Université Laval, 2021. http://hdl.handle.net/20.500.11794/69497.
Pełny tekst źródłaConfidence survey data are time series containting the responses to questions aiming to measure confidence and expectations of economic agents about future economic activity. The richness of these data and their availability in real time attracts the interest of many forecasters who see it as a way to improve their traditional forecasts. In this thesis, I assess the predictive power of survey data for the future evolution of Canadian GDP, while comparing the forecasting performance of the Conference Board of Canada own confidence indices to the indicators I construct using principal component analysis. Using three simple linear models, I carry out an out-of-sample forecasting experiment with rolling windows on the period 1980 to 2019. The results show that principal component analysis provides better-performing indicators than the indices produced by the Conference Board. However, the results of the study cannot show that clear that confidence improves forecasting unambiguently once the lagged growth rate of GDP is added to the analysis.
Alaoui, Ismaili Oumaima. "Clustering prédictif Décrire et prédire simultanément". Thesis, Université Paris-Saclay (ComUE), 2016. http://www.theses.fr/2016SACLA010.
Pełny tekst źródłaPredictive clustering is a new supervised learning framework derived from traditional clustering. This new framework allows to describe and to predict simultaneously. Compared to a classical supervised learning, predictive clsutering algorithms seek to discover the internal structure of the target class in order to use it for predicting the class of new instances.The purpose of this thesis is to look for an interpretable model of predictive clustering. To acheive this objective, we choose to modified traditional K-means algorithm. This new modified version is called predictive K-means. It contains 7 differents steps, each of which can be supervised seperatly from the others. In this thesis, we only deal four steps : 1) data preprocessing, 2) initialization of centers, 3) selecting of the best partition, and 4) importance of features.Our experimental results show that the use of just two supervised steps (data preprocessing and initialization of centers), allow the K-means algorithm to acheive competitive performances with some others predictive clustering algorithms.These results show also that our preprocessing methods can help predictive K-means algorithm to provide results easily comprehensible by users. We are also showing in this thesis that the use of our new measure to evaluate predictive clustering quality, helps our predictive K-means algorithm to find the optimal partition that establishes the best trade-off between description and prediction. It thus allows users to find the different reasons behind the same prediction : two differents instances could have the same predicted label
Durand, Maxim. "Etude de la régulation lymphocytaire T dans deux modèles de transplantation rénale et pulmonaire". Thesis, Nantes, 2018. http://www.theses.fr/2018NANT1001/document.
Pełny tekst źródłaTransplantation is nowadays systematically associated with the administration of immunosuppressive treatments inhibiting the allo-immune response toward the graft to prevent the rejection by the recipient's immune system. However, these heavy treatments have several deleterious effects and are not able to control longterm chronic rejection development. Thus, the objective of this PhD work, based on two distinct clinical situations, is part of two priority axes of transplant research: (1) identify early predictive markers of long-term lung allograft dysfunction and (2) decipher the mechanisms involved in the acceptance of kidney allograft in operational tolerance. We first investigated the mechanisms of graft acceptance in kidney transplanted patients with a functional graft in the absence of immunosuppressive therapy. We identified in these patients a greater proportion of circulating memory regulatory T lymphocytes, with an increased ability to degrade extracellular ATP, a pro-inflammatory mediator, and favour a protolerogenic environment. Thus, this mechanism, not efficient in stable patients under treatments, could participate in the inhibition of the allo-immune response leading to maintaining a functional kidney graft in the absence of immunosuppression. In a second time, we interested in the prediction of chronic rejection occurrence in lung transplanted patients. We report that the early posttransplant regulatory T lymphocyte profile is modified in patients who will develop chronic rejection in the 3 years. Thus, monitoring this predictive biomarker could allow a better identification of patients at risk of rejection
Ghosn, Stéphanie. "Développement d'un modèle animal prédictif de la réponse immunitaire de l'homme aux vaccins". Paris 7, 2014. http://www.theses.fr/2014PA077119.
Pełny tekst źródłaIn the field of vaccine development, a predictive preclinical model is most needed. In this study, our first objective was to compare the validity, for preclinical evaluation of vaccines, of two humanized mouse models, based on the engraftment of NOD-SCID-Gamma null (NSG) mice with either peripheral blood lymphocytes (Hu-PBL-NSG) or human spleen lymphocytes (Hu-SPL-NSG). The second objective was to use the best model to evaluate different vaccine candidates against Plasmodium falciparum malaria in order to select the most promising ones. NSG mice were engrafted with either peripheral blood lymphocytes or spleen lymphocytes and immunized with different vaccine preparations. Our resuits show that the engrafted human lymphocytes remain functional in both models, and secrete IgG, as well as specific antibodies against the injected antigens. However, the Hu-PBL-NSG model exhibits high morbidity and mortality rates. Moreover, human specific antibody responses elicited in this model remain predominantly of the IgM isotype even aliter repeated booster immunizations. These limitations were completely absent in the Hu-SPL-NSG model that has thus been selected to study the human immune response induced by a P. Falciparum liver stage construct LSA3-FL (Liver Stage Antigen 3-Full Length). Our findings corroborate the resuits observed in a clinical trial conducted with LSA3-FL and diverge from those obtained in rodents and primates. In conclusion, we propose that the Hu-SPL-NSG mouse is a cost efficient and time saving preclinical model that allows optimal screening of potentially protective vaccine candidates prior to clinical trials
Luc, Pauline. "Apprentissage autosupervisé de modèles prédictifs de segmentation à partir de vidéos". Thesis, Université Grenoble Alpes (ComUE), 2019. http://www.theses.fr/2019GREAM024/document.
Pełny tekst źródłaPredictive models of the environment hold promise for allowing the transfer of recent reinforcement learning successes to many real-world contexts, by decreasing the number of interactions needed with the real world.Video prediction has been studied in recent years as a particular case of such predictive models, with broad applications in robotics and navigation systems.While RGB frames are easy to acquire and hold a lot of information, they are extremely challenging to predict, and cannot be directly interpreted by downstream applications.Here we introduce the novel tasks of predicting semantic and instance segmentation of future frames.The abstract feature spaces we consider are better suited for recursive prediction and allow us to develop models which convincingly predict segmentations up to half a second into the future.Predictions are more easily interpretable by downstream algorithms and remain rich, spatially detailed and easy to obtain, relying on state-of-the-art segmentation methods.We first focus on the task of semantic segmentation, for which we propose a discriminative approach based on adversarial training.Then, we introduce the novel task of predicting future semantic segmentation, and develop an autoregressive convolutional neural network to address it.Finally, we extend our method to the more challenging problem of predicting future instance segmentation, which additionally segments out individual objects.To deal with a varying number of output labels per image, we develop a predictive model in the space of high-level convolutional image features of the Mask R-CNN instance segmentation model.We are able to produce visually pleasing segmentations at a high resolution for complex scenes involving a large number of instances, and with convincing accuracy up to half a second ahead
Palomares, Maria Lourdes Distor. "La consommation de nourriture chez les poissons : étude comparative, mise au point d'un modèle prédictif et application à l'étude des réseaux trophiques". Toulouse, INPT, 1991. http://www.theses.fr/1991INPT022A.
Pełny tekst źródłaLu, Jun. "Développement de modèles prédictifs décrivant le couplage chimie-transport dans les sédiments naturels". Thesis, Poitiers, 2013. http://www.theses.fr/2013POIT2287/document.
Pełny tekst źródłaManagement of soils or sediments contaminated by metals requires to predict the migration of metallic cations, whose mobility depends both on the transport properties of the medium and chemical reactivity of the system (principally sorption/desorption reactions). To study the sorption of metallic cations (major and trace) in dynamic condition, transport experiments using columns have been carried out with a soil poor in carbonated minerals and organic matter. Considering that the reactivity of this soil was mainly due to swelling clay minerals, a sorption model based on the sorption properties of the Wyoming montmorillonite and built according to a multi-site ion exchanger theory has been integrated into a 1D transport code. The predictions given by this model were then compared with the breakthrough curves measured in this study and those reported in the literature. The study of the reactive transport of major cations highlighted the significant role of protons (even at near neutral pH), and validated the model for major cations (Na and Ca). However, the study concerning Zn (II) showed a discrepancy between the results obtained from batch experiments and those issued from column experiments, which could be attributed to the contribution of another sorbent phase (illite). Finally, the proposed sorption model allowed reproducing with a good confidence experimental data reported from literature for sorption of Zn (II) in dynamic conditions
Zhang, Fan. "Développement d’un dispositif expérimental original et d’un modèle prédictif pour l’étude thermodynamique des composés soufrés". Thesis, Paris, ENMP, 2015. http://www.theses.fr/2015ENMP0069.
Pełny tekst źródłaSulfur compounds are widespread in various industrial fields. Design and/or optimization of the processes involving these compounds require accurate knowledge of thermodynamic properties of the concerned mixtures. Considering the complexity of industrial mixtures and for economical reasons, one may look to thermodynamic models with predictive features. To develop and validate these models, a minimum number of experimental data are required. Our literature review on sulfides (R-S-R'), 1-thiols (R-SH) and their possible solvents (n-alkanes and 1-alkanols) has shown a lack of data and of adapted experimental method. This observation leads to the need of developing a new experimental apparatus.In this work, we designed an innovative experimental apparatus for vapor-liquid-equilibrium measurements in the pressure range of [0.1 – 10] bar. The new apparatus is based on the “static-analytic” method. The key improvements are the two adaptations made for ROLSITM capillary samplers to achieve phase sampling in this pressure range. After validating the newly developed apparatus, new data of binary and ternary mixtures of interest were measured in order to develop a new predictive thermodynamic model focusing on sulfur compounds.The proposed predictive model is the PC-SAFT equation of state combined with a group contribution method (GC-PC-SAFT). We incorporated a polar term into PC-SAFT and investigated two representative families of sulfur compounds: sulfides and 1-thiols. The GC-PC-SAFT model proved reliable in correlating and predicting thermodynamic properties of pure compounds. Model parameters were fitted to the data found in the literature, as well as to those obtained through the new apparatus. Phase diagrams of some binary (sulfide + n-alkane, sulfide + 1-alkanol, 1-thiol + n-alkane and 1-thiol + 1-alkanol) and ternary (1-mercaptan + n-alkane + 1-alcool) mixtures were predicted
Monjo, Florian. "Contrôle prédictif des effets mécaniques de la fatigue musculaire : implication dans l'étude des modèles internes et des modes de coordination entre posture et mouvement". Thesis, Saint-Etienne, 2015. http://www.theses.fr/2015STET014T.
Pełny tekst źródłaMuscle fatigue is a transient and commonly experienced phenomenon. It is mainly associated with loss of force and leads to higher effort to produce a particular force level. The increased discharge rate of the nociceptive afferents (group III and IV) during fatiguing contractions alters motor command expression and finally motor production. An issue that has never been addressed in the literature is the Central Nervous System (CNS) capacity to anticipate muscle fatigue effects in a predictive fashion. This predictive capacity will be investigated thanks to experimental paradigms involving postural predictive processes of control, namely Anticipatory Postural Adjustments (APAs). Because numerous works show that APAs are modulated as a function of the mechanical properties of the upcoming movement, the induction of muscle fatigue at the levels of the focal muscles allowed us to appreciate the CNS capacity to predict muscle fatigue effects. We will demonstrate that this capacity is condition-dependent, i.e. it depends on the nature of the fatiguing contractions performed (voluntary vs. electro-induced) and on the level of cognitive and temporal constraints during movement preparation
Haddadi, Ahmed Zine El Abidine. "Construction d’un score prédictif du risque nosocomial pour des patients de réanimation". Thesis, Lille 2, 2013. http://www.theses.fr/2013LIL2S039/document.
Pełny tekst źródłaLimiting nosocomial infections is still a health challenge although the technical development has improved. They are inherent in medical care and the health care services have the highest prevalence. Indeed, whatever the service (surgical, medical or both), the patients life-giving process is under attack because of the emergence of one or several organ faillures;This generates a diagnostic and therapeutic arsenal which is often invasive.Among the consequences resulting from these infections we will take into account :i) a longer stay in hospitalii) an extra costiii) a higher mortality rateiv) bacterial resistance .If we could anticipate upstream and downstream this issue with complex origins and sometimes fatal consequences, it would be a major asset for patients and a strategic tool for medical teams.The present study is organized in three parts, and first focusses onto the identification of the nosocomial event and death risk factors in intensive care where the study took place. We took into account the the case-mix of the intensive care unit in the TIMONE University Hospital. The study was made with two different statistic methods that is logistic regression and the competitive risks method.The next step first consisted in comparing the predictive capacities of the APACHE II, LOD, SOFA and SAPS II scores in nosocomial patients hospitalized in intensive care . Then it tried to determine if the variation of the LOD, SOFA, APACHEII and SAPS II scores was a prognostic risk factor.Results showed that the best predictive performance was objectively measured by the SOFA and that only the variation of this score between the first day in hospital and the day of the diagnosis of a nosocomial infection, calculated thanks to the AUC, could be predictive of a nosocomal risk. After these steps, and with the results calculated , the construction of a predictive score could be established thanks to the logistic regression method. The objective of this score is to help, or even influence the prescribing doctors when they take decisions or when they try to adjust their therapeutic practices
Kerhascoët, Hugo. "De la mesure du vent au pilotage automatique d'un voilier : modélisation, optimisation & application du traitement du signal prédictif". Thesis, Lorient, 2017. http://www.theses.fr/2017LORIS433.
Pełny tekst źródłaThis thesis is part of NKE Marine’s attempts to improve the quality of apparent and real wind measurementsprovided by its navigation unit. These improvements are expected to benefit both users and improve autopilot performance. The first part describes the specific context of wind measurement on a sailboat, and the inherent issues. This study identified two knotty scientific problems: on-board wind measurements recorded by the sensor (linked to its design and response model) and how the wind triangle is reconstructed. The first objective of this thesis is therefore to propose a new methodology for the design of the wind vane that can match the sensor’s profile with the boat’s characteristics and its route planning. Improvements to this aerial sensor make it possible, on the one hand, to divide the response time by two (via the optimization of the shape of the sensors) and, on the other hand, to provide a useful signal that is twelve times more sensitive (via the application of predictive signal processing techniques that take account of sensor dynamics). Next, previous work on the problem of reconstructing three-dimensional true wind is described in detail. This part begins with a presentation of the relevant standards and definitions of metrics, before examining some solutions that can take account of perturbations in apparent and true wind in the lower layer of the atmosphere. The following part examines current methods for reconstructing true wind. This analysis highlights the fact that these corrections are not applied in the correct frames. To overcome these problems, we propose a new approach to the reconstruction of the true wind, which implements the corrections described above. This manuscript ends with a presentation of the contribution to auto piloting. Specifically, it takes the form of a new concept, termed Super Mode, which can be overlaid onto classical modes of piloting. The aim is to improve safety in gusty conditions (Gust mode), and performance when surfing (Surf mode). The value of this innovation has been proven by competition victories; it has received an innovation award, and Surf mode is the subject of a patent application. The pilot is now able to glide like a human surfer. The final contribution to the autopilot concerns the integration of a new type of sensor: electronic telltales. Finally, this manuscript contributes to our overall understanding of the problems and nuances that must be taken into account when developing or using a navigation unit and an autopilot. The work highlights the close correlation between the various sensors and the data calculated by the navigation unit. Significant improvements to the accuracy of both elements can be obtained when the concept of predictive filtering is extended and appliedto the navigation unit as a whole
Hoang, Tuan Nha. "Incertitude des données biomécaniques : modélisation et propagation dans les modèles de diagnostic des pathologies du système musculosquelettique". Thesis, Compiègne, 2014. http://www.theses.fr/2014COMP2171/document.
Pełny tekst źródłaThe aim of the project is to investigate the modeling of the reliability/incertitude/imprecision of biomedical and biomechanics data (medical images, kinematics/kinetics/EMG data, etc.) and its propagation in the predictive diagnosls models of the disorders of musculoskeletal systems. These diagnosis models will be based on multimodal and multidimensional patient data (3D medical imaging, mechanical data,dinical data,etc.). The literature-based data have been collected to estabish an uncertainty space, which represents fused data from multiple sources, of morphological, mechanical, and movement analysis properties of the musculoskeletal system from multiple sources (i.e. research papers from Science Direct and Pubmed). After that,a new clustering method (US-ECM) is proposed for integrating fused data from multiple sources ln form of a multidimensional uncertainty space (US). Reliability of biomechanical data was evaluated by a fusion approach expert opinion. Reliability criteria of a data source (ie scientific paper published) focus on the technique used the acquisition protocol and measurement and the number of data. A system of questionnaires was developed to co!lect expert opinion. Then, the theory of beliet functions has been applied to merge these opinion to establish a confidence level each data source
Enot, David. "La modélisation moléculaire en tant qu'outil prédictif et de compréhension du comportement thermotrope et lyotrope de sucres. Approche QSPR de la lipophilie de 1,2-dithiole-3-thiones et 1,2-dithiole-3-ones". Rennes 1, 2001. http://www.theses.fr/2001REN10117.
Pełny tekst źródłaHébel, Pascale. "Utilisation d'estimateurs à rétrécisseurs dans les modèles prédictifs ; application à la prédiction du rendement de blé d'hiver". Toulouse 3, 1992. http://www.theses.fr/1992TOU30270.
Pełny tekst źródłaBurnol, Stéphane. "Modèle prédictif d'évolution des Accidents Vasculaires Cérébraux". Thesis, Saint-Etienne, 2010. http://www.theses.fr/2010STET005T.
Pełny tekst źródłaThe lesions resulting from an ischemic stroke remain invisible for several hours or even tens of hours on conventional techniques of medical imaging (CT scan or MRI). New MRI sequences (diffusion-weighted and perfusion-weighted images) reveal the damages immediately. The diffusion and perfusion images reflect complex pathophysiological mechanisms and also contain some information on the fate of the tissue. A probabilistic model of evolution of the lesions visible at an early stage was set up so as to predict expansions and contractions of the damages. The result is presented as a probabilistic mapping of necrosis where the value of each voxel of the image represents the risk of cellular death in the chronic phase. This mapping can then be taken into account in the therapeutic decision. We implemented two types of approach: a segmentation method, allowing extracting regions atrisk, and a prediction method, type of logistic regression, to predict voxels at risk of necrosis. These models provide similar performances with regard to the few models described in the literature, but are more efficient for small lesions and show a big improvement because they are fully automatic
Guenot, Sophie. "Développement d’un modèle prédictif de l’énantiosélectivité des lipases". Toulouse, INSA, 2007. http://eprint.insa-toulouse.fr/archive/00000347/.
Pełny tekst źródłaThe study focused on the development of a predictive model of Burkholderia cepacia ATCC 21808 lipase enantioselectivity and more specifically on the relationship that may exist between enantioselectivity and the accessibility to the active site of (R/S)-alpha-bromo phenyl acetic acid ethyl ester derivatives. A mixed approach, combining molecular modelling techniques and path planning algorithms issued from robotics research, was used to compute the geometrically allowed trajectories of various alpha-bromo phenyl acetic acid ethyl ester derivatives. In silico results indicated that the pathways are computed faster for the (R) enantiomer than the (S) enantiomer for 7 substrates synthesized by chemical synthesis. These results are in qualitative agreement with the enzyme enantioselectivity experimentally determined, indicating that enantioselectivity could be associated with a greater or lesser difficulty of access. In addition, the analysis of trajectories allowed to identify targets for mutagenesis (collision points frequently encountered for all substrates), amongst which three positions were selected: Leu-17, Val-266 and Leu-287. In order to generate and characterize these rational mutants, a recombinant expression system of B. Cepacia lipase in E. Coli and a purification protocol by affinity chromatography have been developed. Genes lip and hp encoding respectively for the lipase and its chaperone protein were cloned in different expression vectors under the control of different promoters. The highest level of lipase activity in soluble form ever described in the literature for this enzyme was obtained by using the vector pFLAG-ATS under the control of the tac promoter (6679 U / Lculture). Rational mutants were constructed by site-directed mutagenesis by replacing each position by the 19 other possible amino acids. Mutants V266, displaying the highest lipase activities in screening with para-nitrophenyl butyrate, were selected to measure their enantioselectivity during hydrolysis of the (R,S)- a-bromo phenyl acetic acid 2-chloro ethyl ester. Among these mutants, V266G displayed an inverted enantioselectivity in favour of the (S)-enantiomer (E=18). The absence of side chain in glycine decreases the steric hindrance at position 266, allowing thus the (S)-enantiomer to (a) access more easily to the catalytic site or (b) adopt a more easily productive conformation in order to form the tetrahedral intermediate. In parallel, a protocol for the production of B. Cepacia lipase at microplate scale has been developed and validated on the library of V266 mutants, in order to dispose of a screening tool for larger mutants libraries created by directed evolution
Zhu, Jian. "Modèles prédictifs de toxicité en radiothérapie par modulation d'intensité". Phd thesis, Université Rennes 1, 2013. http://tel.archives-ouvertes.fr/tel-00855266.
Pełny tekst źródłaZhu, Jian. "Modèles prédictifs de toxicité en radiothérapie par modulation d’intensité". Thesis, Rennes 1, 2013. http://www.theses.fr/2013REN1S017/document.
Pełny tekst źródłaThis thesis is focused on the predictive models of irradiation induced toxicities in intensity modulated radiotherapy. Six different NTCP models were implemented and their parameters were identified at predicting late rectal and bladder toxicities in prostate cancer. Their predictive skills have been demonstrated on both organs. Second, LKB model was used to predict the irradiation induced acute esophagitis after nun-small-cell lung cancer. Then, the benefit of using EUD in prostate cancer IMRT inverse planning was evaluated. The evaluation of the proposed approach proved that the use of EUD significantly decreased both the dose in the bladder and rectum walls. Then, the incorporation of different biological models in IMRT optimization process has been realized. Objective functions were established for different biological factors like NTCP, EUD and TCP. Obtained results show the superiority of the optimization based on biological factors over the optimization relying only on physical factors. Finally, classical NTCP models were corrected to deal with another radiobiological parameter, the α/β ratio. With this additional factor, NTCP models can be extended to predict toxicity for patients with different dose fractionation, these kinds of treatments being more and more clinically used
Chagneau, Pierrette. "Modélisation bayésienne hiérarchique pour la prédiction multivariée de processus spatiaux non gaussiens et processus ponctuels hétérogènes d'intensité liée à une variable prédite : application à la prédiction de la régénération en forêt tropicale humide". Montpellier 2, 2009. http://www.theses.fr/2009MON20157.
Pełny tekst źródłaOne of the weak points of forest dynamics models is the recruitment. Classically, ecologists make the assumption that recruitment mainly depends on both spatial pattern of mature trees and environment. A detailed inventory of the stand and the environmental conditions enabled them to show the effects of these two factors on the local density of seedlings. In practice, such information is not available: only a part of seedlings is sampled and the environment is partially observed. The aim of the paper is to propose an approach in order to predict the spatial distribution and the seedlings genotype on the basis of a reasonable sampling of seedling, mature trees and environmental conditions. The spatial pattern of the seedlings is assumed to be a realization of a marked point process. The intensity of the process is not only related to the seed and pollen dispersal but also to the sapling survival. The sapling survival depends on the environment; so the environment must be predicted on the whole study area. The environment is characterized through spatial variables of different nature and predictions are obtained using a spatial hierarchical model. Unlike the existing models which assume the environmental covariables as exactly known, the recruitment model we propose takes into account the error related to the prediction of the environment. The prediction of seedling recruitment in tropical rainforest in French Guiana illustrates our approach
Michalon, Olivier. "Modèles statistiques pour la prédiction de cadres sémantiques". Thesis, Aix-Marseille, 2017. http://www.theses.fr/2017AIXM0221/document.
Pełny tekst źródłaIn natural language processing, each analysis step has improved the way in which language can be modeled by machines. Another step of analysis still poorly mastered resides in semantic parsing. This type of analysis can provide information which would allow for many advances, such as better human-machine interactions or more reliable translations. There exist several types of meaning representation structures, such as PropBank, AMR and FrameNet. FrameNet corresponds to the frame semantic framework whose theory has been described by Charles Fillmore (1971). In this theory, each prototypical situation and each different elements involved are represented in such a way that two similar situations are represented by the same object, called a semantic frame. The work that we will describe here follows the work already developed for machine prediction of frame semantic representations. We will present four prediction systems, and each one of them allowed to validate another hypothesis on the necessary properties for effective prediction. We will show that semantic parsing can also be improved by providing prediction models with refined information as input of the system, with firstly a syntactic analysis where deep links are made explicit and secondly vectorial representations of the vocabulary learned beforehand
Gimadiev, Timur. "Modèles prédictifs pour les paramètres cinétiques et thermodynamiques des réactions chimiques". Thesis, Strasbourg, 2018. http://www.theses.fr/2018STRAF007/document.
Pełny tekst źródłaThis work is devoted to QSPR modeling of kinetic and thermodynamic properties of chemical reactions using the Condensed Graph of Reaction (CGR) approach. CGR allows encoding structures of reactants and products into one sole molecular graph for which molecular descriptors can be generated. A comprehensive database containing some 11000 manually collected reactions has been developed then used in the modeling. Predictive models were built for rate constants of Diels-Alder, SN2 and E2 reaction as well as for equilibrium constants of tautomeric transformations. They are available for the users via WEB portal. A part of the thesis concerned quantum mechanics studies of reactions between sydnones and strained alkynes for which the size of the dataset was not sufficient to produce statistically meaningful models
Huang, He. "Développement de modèles prédictifs pour la robustesse électromagnétique des composants électroniques". Thesis, Toulouse, INSA, 2015. http://www.theses.fr/2015ISAT0036/document.
Pełny tekst źródłaOne important objective of the electromagnetic compatibility (EMC) studies is to make the products compliant with the EMC requirement of the customers or the standards. However, all the EMC compliance verifications are applied before the delivery of final products. So we might have some new questions about the EMC performance during their lifetime. Will the product still be EMC compliant in several years? Can a product keep the same EMC performance during its whole lifetime? If not, how long the EMC compliance can be maintained? The study of the long-term EMC level, which is called “electromagnetic robustness”, appeared in the recent years. Past works showed that the degradation caused by aging could induce failures of electronic system, including a harmful evolution of electromagnetic compatibility. In this study, the long-term evolution of the EMC levels of two electronic component groups has been studied. The first electronic component type is the integrated circuit. The high-frequency currents and voltages during the switching activities of ICs are responsible for unintentional emissions or coupling. Besides, ICs are also very often the victim of electromagnetic interference. Another group of components is the passive component. In an electronic system, the IC components usually work together with the passive components at PCB level. The functions of passive components in an electronic system, such as filtering and decoupling, also have an important influence on the EMC levels.In order to analyze the long-term evolution of the EMC level of the electronic components, the study in this thesis tends to propose general predictive methods for the electromagnetic compatibility levels of electronic components which evolve with time
Asgari, Jamal. "Etude de modèles prédictifs dans un réseau de stations GPS permanentes". Observatoire de Paris (1667-....), 2005. https://hal.science/tel-02071417.
Pełny tekst źródłaThis thesis consists of two parts with various objectives. The motivations of the first part of the study are : investigation of the GPS spatial and temporal error models, and study of their validity, emphasizing on the network real time applications and the non-differential processing. In the second part we studied the theory of non-differential positioning and we developed a non-differential GPS processing software in the MATLAB IDE. In this study, we investigated the multi-reference use of the network with establishment of interpolation models of the various errors affecting GPS measurements. Filtering and adjustment methods like Kalman filter, collocation, and kriging are studied. The Least Squares Spectral (LSSA) is applied to a 4 years series of TEC values. These analyses show that the data spectrum contains 1 day, 12 hours and 8 hours periods. The validity of several ionospheric models is studied, by examining their effectiveness for real data. Moreover we present a new computation method of ionospheric grids for the prediction of TEC values. The predicted grids could be diffused by Internet and be used in single frequency applications. The numerical tests of interpolation via collocation and kriging were performed for the residual part of the tropospheric delay. The interpolation of other errors may be done in a similar manner. GPSS is a Precise Point Positioning (PPP) software which uses raw data in RINEX format and precise ephemeris. It includes additional visualizations and data quality controls tools
Litalien, David. "Persévérance aux études de doctorat (Ph.D.) : Modèle prédictif des intentions d'abandon". Thesis, Université Laval, 2014. http://www.theses.ulaval.ca/2014/30573/30573.pdf.
Pełny tekst źródłaIn most OECD countries, there is a significant increase in doctoral studies enrollment. Nevertheless, completion rates remain low and are estimated between 40 and 60% for Canadians and Americans. To our knowledge, there is a lack of research on doctoral student’s persistence and available theoretical models to comprehend this phenomenon are difficult to operationalize. Based on self-determination theory (Deci & Ryan, 1985, 2012a, 2012b), the purpose of this dissertation is to provide a general understanding of doctoral studies persistence. It includes two articles. The first article aims to develop and validate a scale to assess five types of regulation (intrinsic, integrated, identified, introjected, and external) toward PhD studies. Results of two studies (N = 244 and N = 1061) support the scale’s reliability, factor structure, and convergent and discriminant validity. The purpose of the second article is to develop and validate a predictive model of dropout intentions. This model posits that perceived competence decreases dropout intentions, and that perceived competence is explained by autonomous and controlled regulations, which are in turn predicted by perceived psychological needs support provided by the student’s advisor, faculties as well as other graduate students. A two-step approach was used: 1) a retrospective comparison of completers and noncompleters (N = 422), and 2) a prospective examination of enrolled PhD students over two semesters to assess dropout intentions (N = 1060). Overall, the findings of the two studies are similar and support the proposed model. Specifically, perceived competence appears to be the cornerstone of doctoral studies persistence intentions, and is predicted mainly by autonomous and controlled regulations and advisor support. Theoretical, methodological, and practical implications are discussed and directions for further research are proposed.
Tableau d'honneur de la FÉSP
Charland, Eric. "Modèles de prédiction de krachs boursiers issus de bulles économiques". Mémoire, Université de Sherbrooke, 2011. http://savoirs.usherbrooke.ca/handle/11143/4901.
Pełny tekst źródłaBonneu, Michel. "Choix de modèles linéaires généralisés en vue de la prédiction". Toulouse 3, 1986. http://www.theses.fr/1986TOU30103.
Pełny tekst źródłaRolland, Chloé. "Modèles orientés objet pour une meilleure prédiction du trafic internet". Paris 6, 2008. http://www.theses.fr/2008PA066657.
Pełny tekst źródłaCortet, Marion. "Construction et validation des modèles de prédiction : étude des utilités". Thesis, Lyon 1, 2015. http://www.theses.fr/2015LYO10197.
Pełny tekst źródłaMedicine asks for prediction. Prediction is needed at different point in the management of a patient. To take the best decision as possible for complementary exams, or therapeutics. Prediction gives an information to the practitioner and the patient, to take a decision. To build these prediction models, we have data bases. The association between clinical or biological data and the outcome probability can be estimated thanks to these data bases. To measure these associations, logistic regression models are used. They are estimated with maximum likelihood method. To evaluate these models, different criteria exist. These criteria quantify adequacy, discrimination capacity, calibration. These models help to take a decision. Prediction errors lead to decision errors. Consequences of these decisions are measurable with utility theory. Therefore, it is a criteria that measure utility of a model that enables us to select the most useful model. Prediction model building is an important point in obstetrics. Indeed, in case of postpartum haemorrhage, it is important to prevent worsening of the clinical situation, and therefore, to identify patient who will worsen fastly. Fibrinogen level was studied as a predictor of severe postpartum haemorrhage. Clinical variables availables at diagnosis of postpartum haemorrhage was then studied. In case of preterm premature rupture of membranes, there is a decision to take, between two choices that may lead to maternal of neonatal morbidity: preterm birth and chorioamnionitis risk with pregnancy continuation. Markers of chorioamnionitis risk may help the practitioners for decision making, by increasing the information. More and more prediction models are developed in all clinical situations. We must be critical before using these models in real life. Their evaluation must take into account their use, and therefore, their utility in case of decision making
Litalien, David. "Persévérance aux études de doctorat (Ph. D.) : modèle prédictif des intentions d'abandon". Doctoral thesis, Université Laval, 2014. http://hdl.handle.net/20.500.11794/25099.
Pełny tekst źródłaMalgré une importante augmentation des effectifs universitaires de 3e cycle dans la plupart des pays de l’OCDE, les taux de diplomation demeurent faibles, estimés entre 40 et 60% pour les canadiens et les américains. À notre connaissance, les recherches qui portent attention à la persévérance aux études de doctorat sont insuffisantes et les modèles théoriques proposés pour étudier ce phénomène sont difficiles à opérationnaliser et par conséquent ont fait l’objet de très peu d’études empiriques. S’appuyant sur la théorie de l’autodétermination (Deci & Ryan, 1985, 2012a, 2012b), cette thèse a pour objectif général d’offrir une meilleure compréhension de la persévérance aux études de doctorat. Celle-ci comporte deux articles. Le premier article vise l’élaboration et la validation de l’Échelle de motivation aux études de doctorat, qui évalue cinq types de motivation (intrinsèque, intégrée, identifiée, introjectée et externe). Les résultats des deux études (N = 244 et N = 1061) soutiennent la structure factorielle de l'échelle, sa fidélité et sa validité convergente et discriminante. Le deuxième article a pour objectif l’élaboration et la validation d’un modèle des intentions d’abandonner les études de doctorat. Le modèle postule que la perception de compétence diminue les intentions d'abandon. Pour sa part, la perception de compétence est expliquée par les motivations autodéterminée et contrôlée qui, en retour, sont prédites par le soutien offert par le directeur de recherche, le corps professoral et les autres étudiants de cycles supérieurs en ce qui a trait aux besoin psychologiques de l’étudiant. Deux études ont été réalisées dans cet article : 1) une étude rétrospective comparative entre des étudiants non persévérants et des étudiants diplômés (N = 422) et 2) une étude prospective évaluant les intentions d’abandon chez des étudiants inscrits sur deux sessions et (N = 1060). Dans l’ensemble, les résultats des deux études convergent et soutiennent le modèle proposé. Plus précisément, la perception de compétence semble être la pierre angulaire de la persévérance aux études de doctorat, étant prédite par les motivations autodéterminée et contrôlée et le soutien du directeur de recherche. Les implications théoriques, méthodologiques et pratiques sont abordées et des pistes de recherches futures sont proposées.
In most OECD countries, there is a significant increase in doctoral studies enrollment. Nevertheless, completion rates remain low and are estimated between 40 and 60% for Canadians and Americans. To our knowledge, there is a lack of research on doctoral student’s persistence and available theoretical models to comprehend this phenomenon are difficult to operationalize. Based on self-determination theory (Deci & Ryan, 1985, 2012a, 2012b), the purpose of this dissertation is to provide a general understanding of doctoral studies persistence. It includes two articles. The first article aims to develop and validate a scale to assess five types of regulation (intrinsic, integrated, identified, introjected, and external) toward PhD studies. Results of two studies (N = 244 and N = 1061) support the scale’s reliability, factor structure, and convergent and discriminant validity. The purpose of the second article is to develop and validate a predictive model of dropout intentions. This model posits that perceived competence decreases dropout intentions, and that perceived competence is explained by autonomous and controlled regulations, which are in turn predicted by perceived psychological needs support provided by the student’s advisor, faculties as well as other graduate students. A two-step approach was used: 1) a retrospective comparison of completers and noncompleters (N = 422), and 2) a prospective examination of enrolled PhD students over two semesters to assess dropout intentions (N = 1060). Overall, the findings of the two studies are similar and support the proposed model. Specifically, perceived competence appears to be the cornerstone of doctoral studies persistence intentions, and is predicted mainly by autonomous and controlled regulations and advisor support. Theoretical, methodological, and practical implications are discussed and directions for further research are proposed.
Cao, Chao. "Compression d'objets 3D représentés par nuages de points". Electronic Thesis or Diss., Institut polytechnique de Paris, 2021. http://www.theses.fr/2021IPPAS015.
Pełny tekst źródłaWith the rapid growth of multimedia content, 3D objects are becoming more and more popular. Most of the time, they are modeled as complex polygonal meshes or dense point clouds, providing immersive experiences in different industrial and consumer multimedia applications. The point cloud, which is easier to acquire than mesh and is widely applicable, has raised many interests in both the academic and commercial worlds.A point cloud is a set of points with different properties such as their geometrical locations and the associated attributes (e.g., color, material properties, etc.). The number of the points within a point cloud can range from a thousand, to constitute simple 3D objects, up to billions, to realistically represent complex 3D scenes. Such huge amounts of data bring great technological challenges in terms of transmission, processing, and storage of point clouds.In recent years, numerous research works focused their efforts on the compression of meshes, while less was addressed for point clouds. We have identified two main approaches in the literature: a purely geometric one based on octree decomposition, and a hybrid one based on both geometry and video coding. The first approach can provide accurate 3D geometry information but contains weak temporal consistency. The second one can efficiently remove the temporal redundancy yet a decrease of geometrical precision can be observed after the projection. Thus, the tradeoff between compression efficiency and accurate prediction needs to be optimized.We focused on exploring the temporal correlations between dynamic dense point clouds. We proposed different approaches to improve the compression performance of the MPEG (Moving Picture Experts Group) V-PCC (Video-based Point Cloud Compression) test model, which provides state-of-the-art compression on dynamic dense point clouds.First, an octree-based adaptive segmentation is proposed to cluster the points with different motion amplitudes into 3D cubes. Then, motion estimation is applied to these cubes using affine transformation. Gains in terms of rate-distortion (RD) performance have been observed in sequences with relatively low motion amplitudes. However, the cost of building an octree for the dense point cloud remains expensive while the resulting octree structures contain poor temporal consistency for the sequences with higher motion amplitudes.An anatomical structure is then proposed to model the motion of the point clouds representing humanoids more inherently. With the help of 2D pose estimation tools, the motion is estimated from 14 anatomical segments using affine transformation.Moreover, we propose a novel solution for color prediction and discuss the residual coding from prediction. It is shown that instead of encoding redundant texture information, it is more valuable to code the residuals, which leads to a better RD performance.Although our contributions have improved the performances of the V-PCC test models, the temporal compression of dynamic point clouds remains a highly challenging task. Due to the limitations of the current acquisition technology, the acquired point clouds can be noisy in both geometry and attribute domains, which makes it challenging to achieve accurate motion estimation. In future studies, the technologies used for 3D meshes may be exploited and adapted to provide temporal-consistent connectivity information between dynamic 3D point clouds
Badr, Georges. "Modèle théorique et outil de simulation pour une meilleure évaluation des claviers logiciels augmentés d'un système de prédiction de mots". Toulouse 3, 2011. http://thesesups.ups-tlse.fr/1549/.
Pełny tekst źródłaPredictive model and simulation tool for a best evaluation of soft keyboard augmented by words prediction list The software keyboards are used to enable text input in mobility and for devices without physical keyboards, such as the new generation of mobile phones. However, these keyboards have several drawbacks such as slowness text entry and fatigue generated for motor impaired users. The solution was to combine software keyboard to lists containing the words likely to continue the word introduced by the user. While these lists, so-called prediction lists, reduce the number of clicks and the number of operations, the speed of user input has decreased. An experiment with an eye tracking system has identified the "strategies" of the user while using and searching a list of words. These results were helpful to refine the prediction models in order to reduce the gap between the performance predicted and the performance actually recorded. Based on observations made during the first experiment, we propose two variants of the use of word prediction list. The first proposes a new way to interact with the list of words and allows maximum use of it. The second evaluates a repositioning of the list of words in order to reduce the number of eye movements to the list. These two propositions were theoretically and experimentally evaluated by users. These software can improve the input performances compared with a classic word prediction list
Legrand, Mathias. "Modèles de prédiction de l'interaction rotor/stator dans un moteur d'avion". Phd thesis, Ecole centrale de nantes - ECN, 2005. http://tel.archives-ouvertes.fr/tel-00011631.
Pełny tekst źródłaJungman, Elsa. "Développement d'un modèle prédictif de la pénétration percutanée par approches chromatographiques et spectroscopiques". Phd thesis, Université Paris Sud - Paris XI, 2012. http://tel.archives-ouvertes.fr/tel-00819391.
Pełny tekst źródłaRaad, Ali. "Modèle prédictif à base d'approches qualitatives pour la contribution à la radiothérapie adaptative". Thesis, Lille 1, 2015. http://www.theses.fr/2015LIL10166.
Pełny tekst źródłaThe thesis subject is defined in the framework of the fight against cancer. External radiation therapy is a technical treatment of tumor targets by radioactive particles, under guidance of an imaging system. The effectiveness of this macroscopic targeting treatment depends in part on the evolution and movement of the biological target during the treatment, which is planned over several sessions. In our work, we focused to the temporal evolution of the biological target session by session. The goal is to estimate the space covered by the target movement in each session, and to predict the future position of the target and thus to correct or adapt the treatment plan. Our microscopic biological target in the case of head and neck cancer is described by the parotid gland which is sensitive to the radiation. The objective is to minimize avoid radiating the parotid, when the tumor is in its periphery. We proposed two 3D reconstruction methods of the parotid, namely B-Spline Bezier bi-cubic and Marching cubes. Once the reconstruction is performed, the volume of the parotid gland, at each session of treatment is calculated. A safety zone around the parotid gland was defined. Finally, a classification of the patients was performed based on the associated dose for the treatment.At a macroscopic level, we have developed a decision-support approach to refine the relative position of the patient on the treatment table. This approach is based on a modeling of Neural Network and especially MLP and RBF. The performance between two neuronal approximation techniques was analyzed, according to the geometrical positioning of patients
Ulvé, Ronan. "Caractérisation moléculaire et cellulaire des lymphomes canins : modèles précliniques prédictifs des lymphomes homologues humains". Thesis, Rennes 1, 2016. http://www.theses.fr/2016REN1B045.
Pełny tekst źródłaLymphomas are among the most common cancers in humans and dogs. They show strong clinical, histological and response homologies to treatments. In the current context, new generation sequencing (NGS) methods allow identification of many genetic alterations needed for the diagnosis, prognosis and development of targeted therapies. However, the development of new molecules encounters a high proportion of failure in clinical studies. This finding is due in part to the use of models that are not reflect all aspects of the disease occurring in humans. In dogs, artificial selection done by humans means that today, many breeds have predispositions to lymphomas and even to certain subtypes. This characteristic makes the dog a relevant spontaneous model both for the study of the genetic basis of lymphomas and for the development of new molecules for humans with veterinary clinical trials. My thesis work consisted in the genetic characterization of canine lymphomas to propose predictive models of human homologous lymphomas. Following a collection step of a large number of lymphomas cases, I worked on the improvement of a diagnostic test to subtype B or T lymphomas based on amplification called PARR. I also showed a familial transmission of lymphomas in the Bernese Mountain Dog, which allows me to perform a genome-wide association study (GWAS) comprising 63 affected dogs and 167 healthy dogs. I identified several loci on chromosomes 9, 15 and 23, the last one including the MYD88 gene known to be involved in human lymphomas. I have also discovered by different NGS approaches (RNA-Seq and Capture targeted) recurrent genetic alterations shared between the Man and the dog. Among these, I have identified gene fusions between immunoglobulins and cyclins D: 3 cases for CCND3 and 1 case for CCND1. I also found strong recurrences of alterations involving the oncogenes KDR, MYC or UBR5 as well as the tumor suppressor genes POT1, PTEN or TP53. Since these events are associated with aggressive or resistant lymphomas in humans, canine lymphomas are thus of major interest as a spontaneous model. Finally, I have carried out in vitro tests of molecules, which can be carried out from the CLBL-1 cell line or from primary cell cultures characterized by NGS. This preliminary step allows us to consider veterinary clinical trials with owners dogs with lymphomas. This approach is part of the "One Health" concept, which aims to bring this research to human and veterinary medicine
Khanjari, Miyaneh Eynollah. "Un cadre générique pour les modèles globaux fondés sur les motifs locaux". Thesis, Tours, 2009. http://www.theses.fr/2009TOUR4020/document.
Pełny tekst źródłaThe construction of global models is a significant field of Knowledge Discovery in Databases. In particular, global models based on local patterns such as association rules provide a succinct and understandable description of data. The numerous viewpoints, aims and domain-specific data require a wide range of global models and associated construction methods. This thesis proposes a generic framework for formalizing and manipulating global models based on local patterns. In this framework, a lot of the existing construction methods dedicated to classification, clustering and summarization are easily formulated in a declarative way. We provide a generic algorithm enabling to leave aside technical aspects, for instance the kind of used patterns and associated mining approach. Moreover, we also optimize this algorithm according to the specified parameters. Finally, our framework facilitates the comparison of existing construction methods by highlighting their main features
Marchesseau, Stéphanie. "Simulation de modèles personnalisés du coeur pour la prédiction de thérapies cardiaques". Phd thesis, Ecole Nationale Supérieure des Mines de Paris, 2013. http://pastel.archives-ouvertes.fr/pastel-00820082.
Pełny tekst źródłaMarchesseau, Stephanie. "Simulation de modèles personnalisés du coeur pour la prédiction de thérapies cardiaques". Thesis, Paris, ENMP, 2013. http://www.theses.fr/2012ENMP0082/document.
Pełny tekst źródłaThe clinical understanding and treatment of cardiovascular diseases is highly complex. For each patient, cardiologists face issues in determining the pathology, choosing a therapy or selecting suitable patients for the therapy. In order to provide additional guidance to cardiologists, many research groups are investigating the possibility to plan such therapies based on biophysical models of the heart. The hypothesis is that one may combine anatomical and functional data to build patient-specific cardiac models that could have the potential to predict the benefits of different therapies. Cardiac electromechanical simulations are based on computational models that can represent the heart geometry, motion and electrophysiology patterns during a cardiac cycle with sufficient accuracy. Integration of anatomical, mechanical and electrophysiological information for a given subject is essential to build such models.In this thesis, we first introduce the geometry, kinematics and electrophysiology personalizations that are necessary inputs to mechanical modeling. We propose to use the Bestel-Cl'ement-Sorine electromechanical model of the heart, which is sufficiently accurate without being over-parametrized for the available data. We start by presenting a new implementation of this model in an efficient opensource framework for interactive medical simulation and we analyze the resulting simulations through a complete sensitivity analysis.In a second step, the goal is to personalize the mechanical parameters from medical images (MRI data). To this end, we first propose an automatic calibration algorithm that estimates global mechanical parameters from volume and pressure curves. This technique was tested on 7 volunteers and 2 heart failure cases and allowed to perform a preliminary specificity study that intends to determine the relevant parameters able to differentiate the pathological cases from the control cases.Once initialized with the calibrated values, the parameters are then locally personalized with a more complex optimization algorithm. Reduced Order Unscented Kalman Filtering is used to estimate the contractilities on all of the AHA zones of the Left Ventricle, matching the regional volumes extracted from cine MRI data. This personalization strategy was validated and tested on several pathological and healthy cases. These contributions have led to promising results through this thesis and some are already used for various research studies
Coste, Nicolas. "Vers la prédiction de performance de modèles compositionnels dans les architectures GALS". Phd thesis, Grenoble, 2010. http://www.theses.fr/2010GRENM028.
Pełny tekst źródłaValidation, comprising functional verification and performance evaluation, is critical for complex hardware designs. Indeed, due to the high level of parallelism in modern designs, a functionally verified design may not meet its performance specifications. In addition, the later a design error is identified, the greater its cost. Thus, validation of designs should start as early as possible. This thesis proposes a compositional modeling framework, taking into account functional and time aspects of hardware systems, and defines a performance evaluation approach to analyze constructed models. The modeling framework, called Interactive Probabilistic Chain (IPC), is a discrete-time process algebra, representing delays as probabilistic phase type distributions. We defined a branching bisimulation and proved that it is a congruence with respect to parallel composition, a crucial property for compositional modeling. IPCs can be considered as a transposition of Interactive Markov Chains in a discrete-time setting, allowing a precise and compact modeling of fixed hardware delays. For performance evaluation, a fully specified IPC is transformed, assuming urgency of actions, into a discrete-time Markov chain, which can then be analyzed. Additionally, we defined a performance measure, called latency, and provided an algorithm to compute its long-run average distribution. The modeling approach and the computation of latency distributions have been implemented in a tool-chain relying on the CADP toolbox. Using this tool-chain, we studied communication aspects of an industrial hardware design, the xSTtream architecture, developed at STMicroelectronics
Coste, Nicolas. "Vers la prédiction de performance de modèles compositionnels dans les architectures GALS". Phd thesis, Université de Grenoble, 2010. http://tel.archives-ouvertes.fr/tel-00538425.
Pełny tekst źródłaEl, Koujok Mohamed. "Contribution au pronostic industriel : intégration de la confiance à un modèle prédictif neuro-flou". Phd thesis, Université de Franche-Comté, 2010. http://tel.archives-ouvertes.fr/tel-00542230.
Pełny tekst źródłaDzeutouo, Zapa Donard. "Développement d’un modèle prédictif de la productivité spatio-temporelle des plants de bleuets sauvages". Mémoire, Université de Sherbrooke, 2017. http://hdl.handle.net/11143/11331.
Pełny tekst źródłaGiraud, Laurent. "Modèle prédictif pour le développement d'un système embarqué de vidange des lubrifiants pour transmission". Thesis, National Library of Canada = Bibliothèque nationale du Canada, 2000. http://www.collectionscanada.ca/obj/s4/f2/dsk1/tape4/PQDD_0021/NQ48980.pdf.
Pełny tekst źródłaEl, Koujoc Mohamed. "Contribution au pronostic industriel : intégration de la confiance à un modèle prédictif neuro-flou". Besançon, 2010. http://www.theses.fr/2010BESA2024.
Pełny tekst źródłaIndustrial prognostic is nowadays recognized as a key feature to reinforce the dependability of equipments while reducing maintenance costs. However, it can be dfficult to implement an efficient prognostic tool since the lack of knowledge on the behaviour of an equipment can impede the development of classical dependability analysis. In this context, the general purpose of the work is to propose a prognostic system that starts from monitoring data and goes through useful indicators to optimize maintenance strategies. The work also aims at mitigating some problems that follow from the lack of knowledge on the degradation phenomena (amount of data, expertise in building a model). Developments are based on the use of the neuro-fuzzy system exTS as a tool to predict the state of degradation of an equipment. Its structure is partially determined, on one side, thanks to its evolving capability, and on the other side, thanks to parsimony principle : a procedure to automatically generate an accurate exTS prediction system that reaches a compromise between complexity and generalization capability is proposed. Following that, a method to firstly, a priori estimate the probability density function (pdf) of the error of prediction of the neuro-fuzzy system, and secondly propagate it to any prediction step, is also proposed and illustrated. This contribution enables to provide a confidence measure on predictions and thereby to integrate uncertainty to the prognostic process. Finally, mechanisms of reliability evaluation are adapted to the predictive case in order to generate the prognostic metrics that allows optimizing maintenance strategies, notably the Remaining Useful Life (RUL)
Cerrada-Lozada, Mariela. "Sur les modèles flous adaptatifs dynamiques". Toulouse, INSA, 2003. http://www.theses.fr/2003ISAT0021.
Pełny tekst źródłaThis work deals with the proposition of an adaptive fuzzy model with dynamical membership functions. The identification of the parameters of these membership functions is performed by a on-line reinforcement learning-based algorithm. This approach takes into account the system variables dynamic by incorporating the mean value and the variance, at time t, of the input and output variables of the fuzzy model into its membership functions. By this way, the fuzzy sets associated to the fuzzy variables, are relocated on the domain of discourse according to the sampled mean and variance values; thus, a disjointed partition of the fuzzy sets of the fuzzy model could be avoid. The dynamical property of the proposed fuzzy models is an asset in fuzzy control problems in case of time-varying nonlinear systems, for example. Classical examples related to the identification of time-varying nonlinear functions show the capabilities of the dynamical fuzzy models. An application to predictive control has been developed using the fuzzy model as one step ahead predictor and the reinforcement learning in the optimization problem of this type of control scheme. Finally, a discussion about the use of the information provided by the dynamical membership functions is presented in order to accomplish diagnosis and supervision tasks at upper levels
Sbidian, Émilie. "Evolution et facteurs pronostiques de la Neurofibromatose 1". Thesis, Paris 11, 2012. http://www.theses.fr/2012PA11T064/document.
Pełny tekst źródłaNeurofibromatosis-1 (NF1) is a common autosomal dominant condition which is a source of various multisystemic manifestations related either to the accumulation of neurofibromas or to specific developmental abnormalities. There are no obvious factors that predict disease progression. Thus, the aim of our project was to characterize the phenotype of NF1 patients with a severe prognosis. Patients were identified among adults with NF-1 followed up in the Réseau NF-France. The Réseau NF-France is a French medical network devoted to neurofibromatosis 1. It has elaborated recommendations for the management of the disease and recommended a coordinated follow-up in specialized multidisciplinary centres. About 2 500 patients were enrolled. We first evaluated the mortality in a large retrospective cohort of NF1 patients. The standardized mortality ratio (SMR) with its 95% confidence interval (CI) was calculated as the ratio of observed over expected numbers of deaths. Between 1980 and 2006, 1895 NF1 patients were seen. The excess mortality occurred among patients aged 10 to 20 years (SMR=5.2; CI, 2.6-9.3; P<10-4) and 20 to 40 years (SMR, 4.1; 2.8-5.8; P<10-4). The main cause of death was the malignant tumors of the nerve sheath (MPNSTs) developing from preexisting internal neurofibromas. Then, a case-control study including 208 patients with NF1 allowed us to explain the increased risk of mortality among NF1 patients harboring subcutaneous neurofibromas (SC-NF) by the presence of internal neurofibromas (NF) at risk of MPNSTs systematically investigated with imaging (MRI) (OR=4.3, IC95% : 2.2 – 8.2). The association with SC-NF was stronger for patients with ten or more SC-NFs (OR=82, IC95% : 10.4 – 647.9) and for diffuse (OR=14.7, IC95% : 3.8 – 57.3), and ≥ 3 cm (OR=6.3, IC95% : 2.3 – 17.4) internal neurofibromas. Patients with SC-NF constituted 20 to 30% of the NF1 population. So, to characterize patients at risk of developping MPNSTs, we developped and validated a clinical score for predicting internal neurofibromas in adults. Four variables were independently associated with internal neurofibromas: at least two subcutaneous neurofibromas (OR=4.7, IC95% : 2.1 – 10.5), age ≤30 years (OR=3.1, IC95% : 1.4 – 6.8), absence of cutaneous neurofibromas (OR=2.6, IC95% : 0.9 – 7.5), and fewer than six café-au-lait spots (OR=2, IC95% : 0.9 – 4.6). The NF1Score was computed as 10 . [age ≤30 years] + 10 • [absence of cutaneous neurofibromas] + 15 • [≥2 subcutaneous neurofibromas] + 5 • [<6 café-au-lait spots]). Calibration was excellent (Hosmer-Lemeshow statistic=4.53; degrees of freedom=7; P>0.5) and discrimination was good (AUC-ROC= 0.75; 95%CI, 0.7-0.8). Finally clinical expressivity is variable and manifestations of NF1 change at different times in an individual’s life. Consequently, a specific study was needed in pediatric patients. We identified easily recognizable clinical characteristics associated with internal neurofibromas in children with NF1. By multivariate analysis, age (OR=1.1, IC95% : 1.0 – 1.2), xanthogranulomas (OR=4.5, IC95% : 0.9 – 21.7), and presence of both subcutaneous and plexiform neurofibromas (OR=5.0, IC95% : 1.8 – 13.6) were independently associated with internal neurofibromas. Moreover internal neurofibromas increased during adolescence. Excess risk of developing internal neurofibromas seems to occur between the adolescence and the age of to 30 in NF1 patients. These clinical features in adults and children would define a new population at risk for complications that may need closer clinical and imaging follow-up
Labiadh, Mouna. "Méthodologie de construction de modèles adaptatifs pour la simulation énergétique des bâtiments". Thesis, Lyon, 2021. http://www.theses.fr/2021LYSE1158.
Pełny tekst źródłaPredictive modeling of energy consumption in buildings is essential for intelligent control and efficient planning of energy networks. One way to perform predictive modeling is through machine learning approaches. Alongside their good performance, these approaches are time efficient and facilitates the integration of buildings into smart environments. However, accurate machine learning models rely heavily on collecting relevant building operational data in a sufficient amount, notably when deep learning is used. In the field of buildings energy, historical data are not available for training, such is the case in newly built or newly renovated buildings. Moreover, it is common to verify the energy efficiency of buildings before construction or renovation. For such cases, only a contextual description about the future building and its design is available. The goal of this dissertation is to address the predictive modeling tasks of building energy consumption when no historical data are available for the given target building. To that end, existing data collected from multiple different source buildings are leveraged. This is increasingly relevant with the growth of open data initiatives in various sectors, namely building energy. The main idea is to transfer knowledge across building models. There is little research at the intersection of building energy modeling and knowledge transfer. An important challenge arises when dealing with multi-source data, since large domain shift may exist between different sources and also between each source and the target. As a contribution, a two-fold query-adaptive methodology is developed for cross-building predictive modeling. The first process recommends relevant training data to a target building solely by using a minimal contextual description on it (metadata). Contextual descriptions are provided as user queries. To enable a task-specific recommendation, a deep similarity learning framework is used. The second process trains multiple predictive models based on recommended training data. These models are combined together using an ensemble learning framework to ensure a robust performance. The implementation of the proposed methodology is based on microservices. Logically independent workflows are modeled as microservices with single purposes and separate data sources. Building metadata and time series data collected from multiple sources are integrated into an unified ontology-based view. Experimental evaluation of the predictive model factory validates the effectiveness and the applicability for the use case of building energy modeling. Moreover, because of its generic design, the methodology for query-adaptive cross-domain predictive modeling can be re-used for a diverse range of use cases in different fields
Fonte, Christophe. "Une nouvelle régulation pour les processus industriels : la commande adaptative à multiples modèles de référence". Lille 1, 1986. http://www.theses.fr/1986LIL10102.
Pełny tekst źródłaGrenier, Alain Guy. "La valeur prédictive du modèle polyarchique de Robert A Dahl". Thesis, University of Ottawa (Canada), 1988. http://hdl.handle.net/10393/22122.
Pełny tekst źródłaYoudjari, Djonkamla. "Développement d'un modèle de prédiction de l'uni des chaussées flexibles". Doctoral thesis, Université Laval, 2018. http://hdl.handle.net/20.500.11794/31592.
Pełny tekst źródłaThe objective of the thesis is to develop a model of prediction of the flexible pavements roughness. This goal was achieved through a rigorous process that involved four steps. First, a review of knowledge was made. It appears from this study that the International Roughness Index (IRI) is the index that better represents the roughness phenomenon, but the prediction is still empirical. Second, the two main pavement degradations that could cause roughness problems during the useful life are the permanent deformation and frost heave. After analysis of the existing models the need of the prediction model was established. The desired prediction models should have in their mathematical structure the geotechnical parameters of soil, which are the origin of variability along a road. For frost heave, the Konrad model (1981 and 2005) was found to be appropriate. However, for permanent deformation there was no adequate model. Thus, a permanent deformation model was developed. Third, a mathematical model for the prediction of pavements roughness in terms of IRI was derived. It has been built on rigorous assumptions and includes the following parameters: permanent deformation ( ε p ), frost heave (S), transverse crack ( Ft ) and coefficient representing the effect of amplification or attenuation of wavelengths (K). Fourth, the derived model was calibrated and validated with data obtained on actual pavements. For calibration, data from the LTTP and MTQ databases and the geotechnical parameters resulting from the characterization tests of the samples, taken at every 5 m along 5 different sites in Quebec Province were used. The required calibration coefficients have been determined successfully. For the validation of the model, due to lack of geotechnical parameters at every 5 m along the identified sections to achieve this objective, the principle of the model use level has been developed. Three levels of the model use were developed and the model was validated at the third level with an acceptable level of success.