Estimation du paramètre d'une copule archimedienne en présence de censure dépendante

Les méthodes classiques d’analyse de survie notamment la méthode non paramétrique de Kaplan et Meier (1958) supposent l’indépendance entre les variables d’intérêt et de censure. Mais, cette hypothèse d’indépendance n’étant pas toujours soutenable, plusieurs auteurs ont élaboré des méthodes pour prendre en compte la dépendance. La plupart de ces méthodes émettent des hypothèses sur cette dépendance. Dans ce mémoire, nous avons proposé une méthode d’estimation de la dépendance en présence de censure dépendante qui utilise le copula-graphic estimator pour les copules archimédiennes (Rivest etWells, 2001) et suppose la connaissance de la distribution de la variable de censure. Nous avons ensuite étudié la consistance de cet estimateur à travers des simulations avant de l’appliquer sur un jeu de données réelles.

Techniques for batch reinforcement learning in robotics

This thesis addresses the Batch Reinforcement Learning methods in Robotics. This sub-class of Reinforcement Learning has shown promising results and has been the focus of recent research. Three contributions are proposed that aim to extend the state-of-art methods allowing for a faster and more stable learning process, such as required for learning in Robotics. The Q-learning update-rule is widely applied, since it allows to learn without the presence of a model of the environment. However, this update-rule is transition-based and does not take advantage of the underlying episodic structure of collected batch of interactions. The Q-Batch update-rule is proposed in this thesis, to process experiencies along the trajectories collected in the interaction phase. This allows a faster propagation of obtained rewards and penalties, resulting in faster and more robust learning. Non-parametric function approximations are explored, such as Gaussian Processes. This type of approximators allows to encode prior knowledge about the latent function, in the form of kernels, providing a higher level of exibility and accuracy. The application of Gaussian Processes in Batch Reinforcement Learning presented a higher performance in learning tasks than other function approximations used in the literature. Lastly, in order to extract more information from the experiences collected by the agent, model-learning techniques are incorporated to learn the system dynamics. In this way, it is possible to augment the set of collected experiences with experiences generated through planning using the learned models. Experiments were carried out mainly in simulation, with some tests carried out in a physical robotic platform. The obtained results show that the proposed approaches are able to outperform the classical Fitted Q Iteration.

In situ characterization of sediment oxygen demand across the spectrum of non-resuspension to full resuspension with varying bed shear stress

Sediment oxygen demand (SOD) can be a significant oxygen sink in various types of water bodies, particularly slow-moving waters with substantial organic sediment accumulation. In most settings where SOD is a concern, the prevailing hydraulic conditions are such that the impact of sediment resuspension on SOD is not considered. However, in the case of Bubbly Creek in Chicago, Illinois, the prevailing slack water conditions are interrupted by infrequent intervals of very high flow rates associated with pumped combined sewer overflow (CSO) during intense hydrologic events. These events can cause resuspension of the highly organic, nutrient-rich bottom sediments, resulting in precipitous drawdown of dissolved oxygen (DO) in the water column. While many past studies have addressed the dependence of SOD on near-bed velocity and bed shear stress prior to the point of sediment resuspension, there has been limited research that has attempted to characterize the complex and dynamic phenomenon of resuspended-sediment oxygen demand. To address this issue, a new in situ experimental apparatus referred to as the U of I Hydrodynamic SOD Sampler was designed to achieve a broad range of velocities and associated bed shear stresses. This allowed SOD to be analyzed across the spectrum of no sediment resuspension associated with low velocity/ bed shear stress through full sediment resuspension associated with high velocity / bed shear stress. The current study split SOD into two separate components: (1) SODNR is the sediment oxygen demand associated with non-resuspension conditions and is a surface sink calculated using traditional methods to yield a value with units (g/m2/day); and (2) SODR is the oxygen demand associated with resuspension conditions, which is a volumetric sink most accurately characterized using non-traditional methods and units that reflect suspension in the water column (mg/L/day). In the case of resuspension, the suspended sediment concentration was analyzed as a function of bed shear stress, and a formulation was developed to characterize SODR as a function of suspended sediment concentration in a form similar to first-order biochemical oxygen demand (BOD) kinetics with Monod DO term. The results obtained are intended to be implemented into a numerical model containing hydrodynamic, sediment transport, and water quality components to yield oxygen demand varying in both space and time for specific flow events. Such implementation will allow evaluation of proposed Bubbly Creek water quality improvement alternatives which take into account the impact of SOD under various flow conditions. Although the findings were based on experiments specific to the conditions in Bubbly Creek, the techniques and formulations developed in this study should be applicable to similar sites.

Signs and Symptoms of Dry Eye in Keratoconus Patients: A Pilot Study

Purpose: To compare signs and symptoms of dry eye in keratoconus (KC) patients versus healthy subjects. Methods: A total of 15 KC patients (KC group, n = 15 eyes) and 16 healthy subjects (control group, 16 eyes) were enrolled in this study. The Schirmer I test with no anesthetic, tear break-up time (TBUT), corneal staining characteristics, and ocular surface disease index (OSDI) scores were evaluated for both groups. Impression cytology, combined with/scanning laser confocal microscopy (LCM), was performed to evaluate goblet cell density, mucin cloud height (MCH), and goblet cell layer thickness (CLT). Finally, tear concentrations of di-adenosine tetraphosphate (Ap4A) were assessed. Results were statistically analyzed using Shapiro–Wilk and non-parametric Wilcoxon rank sum tests. Statistical significance was set at p < 0.05. Results: KC patients had lower tear volumes and greater corneal staining than did healthy subjects (p < 0.05). OSDI scores were 44.96 ± 8.65 and 17.78 ± 6.50 for the KC and control groups, respectively (p < 0.05). We found no statistically significant differences in TBUT between groups. Impression cytology revealed lower goblet cell densities in KC group patients versus control group subjects (84.88 ± 32.98 and 128.88 ± 50.60 cells/mm,2 respectively, p < 0.05). There was a statistically significant reduction in MCH and CLT in KC group patients compared with control group subjects. Ap4A tear concentrations were higher in KC group patients than in control group subjects (2.56 ± 1.10 and 0.15 ± 0.12 µM, respectively, p < 0.05). Conclusions: The parameters evaluated in this study indicate that KC patients suffer greater symptoms of dry eye and greater tear instability, primarily due to the decreased mucin production in their tears, than do healthy patients with no KC.

Effects of acute exercise on salivary free insulin-like growth factor 1 and interleukin 10 in sportsmen.

Background: Saliva analysis is rapidly developing as a tool for the assessment of biomarkers of sports training. It remains poorly understood whether a short bout of sport training can alter some salivary immune biomarkers. Aim: To investigate the effect of acute exercise using football training session on salivary flow rate, salivary free Insulin-like Growth Factor-1 (IGF-1) and Interleukin 10 (IL-10). Methods: Saliva samples were collected before and immediately after a football session. Salivary flow rates, salivary levels of free IGF-1 and IL-10 (using ELISA) were determined. Data was analyzed and compared using Related Samples Wilcoxon Signed Rank test (non-parametric test). Relationships between salivary flow rate and levels of free IGF-1 and IL-10 were determined using Spearman correlation test. Results: There were 22 male footballers with a mean age of 20.46 years. Salivary flow rate reduced significantly (p = 0.01) after the training session while salivary levels of free IGF-1 and IL-10 did not show any significant change. Also, there were no correlations between salivary flow rates and salivary levels of free IGF-1 and IL-10 before and after exercise. Conclusion: These findings suggest that acute exercise caused significant reduction in salivary flow rate but no change in the levels of salivary free IGF-1 and IL-10.

Active Foreground Region Extraction and Tracking for Sports Video Annotation

Automatic video segmentation plays a vital role in sports videos annotation. This paper presents a fully automatic and computationally efficient algorithm for analysis of sports videos. Various methods of automatic shot boundary detection have been proposed to perform automatic video segmentation. These investigations mainly concentrate on detecting fades and dissolves for fast processing of the entire video scene without providing any additional feedback on object relativity within the shots. The goal of the proposed method is to identify regions that perform certain activities in a scene. The model uses some low-level feature video processing algorithms to extract the shot boundaries from a video scene and to identify dominant colours within these boundaries. An object classification method is used for clustering the seed distributions of the dominant colours to homogeneous regions. Using a simple tracking method a classification of these regions to active or static is performed. The efficiency of the proposed framework is demonstrated over a standard video benchmark with numerous types of sport events and the experimental results show that our algorithm can be used with high accuracy for automatic annotation of active regions for sport videos.

Os efeitos das reclamações online na lealdade dos consumidores: um estudo experimental

In an environment of constant change, technological developments, market competition and more informed consumers, the search for a lasting relationship through the conquest of loyalty has become the objective of companies. However, several authors suggest that this loyalty can be affected by negative comments available on the internet. Therefore, this dissertation has as objective to examine if the complaints are available on the internet impact the loyalty to a brand of mobile phone. The research used as the basis the Expanded NCSB model suggest by Johnson et al. (2001), studying five prominent drives of loyalty: image/brand reputation, affective commitment, calculative commitment, perceived value and trust, beyond the satisfaction construct as moderator variable. The research method adopted was the experimental design which included 285 undergraduate students, with the trial which included 285 undergraduate students, with the field study of the mobile industry, specifically, the brands of cell phones. The research approach was quantitative and methods were descriptive statistics, factor analysis, cluster analysis, linear regression and non-parametric test of Wilcoxon for data analysis. Of the 16 hypothesis stemmed from the research model proposed, 12 were confirmed. The results showed that the complaint available on the internet, here represented by the available on the site Reclame Aqui, may impact consumer perceptions about brand loyalty, as well as its antecedents, being that these complaints can affect all the consumers, regardless of historical satisfaction with the brand. It also noted the positive relationship between the independent variables trust, image/brand reputation, perceived value, affective commitment and calculative commitment and the dependent variable - loyalty, even when considering the data obtained after exposure to the complaint. However, no unanimous conclusion that the relationship between these variables was strongest in the group with satisfactory experience. At the first moment of the research, the trust was the most important variable for the formation of loyalty. However, after exposure to treatment, the image/brand reputation, was more relevant. Contributions of the study, limitations and recommendations for future researches are approached in the present investigation

Outcome Prediction of Consciousness Disorders in the Acute Stage Based on a Complementary Motor Behavioural Tool.

INTRODUCTION: Attaining an accurate diagnosis in the acute phase for severely brain-damaged patients presenting Disorders of Consciousness (DOC) is crucial for prognostic validity; such a diagnosis determines further medical management, in terms of therapeutic choices and end-of-life decisions. However, DOC evaluation based on validated scales, such as the Revised Coma Recovery Scale (CRS-R), can lead to an underestimation of consciousness and to frequent misdiagnoses particularly in cases of cognitive motor dissociation due to other aetiologies. The purpose of this study is to determine the clinical signs that lead to a more accurate consciousness assessment allowing more reliable outcome prediction. METHODS: From the Unit of Acute Neurorehabilitation (University Hospital, Lausanne, Switzerland) between 2011 and 2014, we enrolled 33 DOC patients with a DOC diagnosis according to the CRS-R that had been established within 28 days of brain damage. The first CRS-R assessment established the initial diagnosis of Unresponsive Wakefulness Syndrome (UWS) in 20 patients and a Minimally Consciousness State (MCS) in the remaining13 patients. We clinically evaluated the patients over time using the CRS-R scale and concurrently from the beginning with complementary clinical items of a new observational Motor Behaviour Tool (MBT). Primary endpoint was outcome at unit discharge distinguishing two main classes of patients (DOC patients having emerged from DOC and those remaining in DOC) and 6 subclasses detailing the outcome of UWS and MCS patients, respectively. Based on CRS-R and MBT scores assessed separately and jointly, statistical testing was performed in the acute phase using a non-parametric Mann-Whitney U test; longitudinal CRS-R data were modelled with a Generalized Linear Model. RESULTS: Fifty-five per cent of the UWS patients and 77% of the MCS patients had emerged from DOC. First, statistical prediction of the first CRS-R scores did not permit outcome differentiation between classes; longitudinal regression modelling of the CRS-R data identified distinct outcome evolution, but not earlier than 19 days. Second, the MBT yielded a significant outcome predictability in the acute phase (p<0.02, sensitivity>0.81). Third, a statistical comparison of the CRS-R subscales weighted by MBT became significantly predictive for DOC outcome (p<0.02). DISCUSSION: The association of MBT and CRS-R scoring improves significantly the evaluation of consciousness and the predictability of outcome in the acute phase. Subtle motor behaviour assessment provides accurate insight into the amount and the content of consciousness even in the case of cognitive motor dissociation.

Comparing the performance of the EQ-5D-3L and the EQ-5D-5L in young Portuguese adults

Background Some studies have reported a ceiling effect in EQ-5D-3L, especially in healthy and/or young individuals. Recently, two further levels have been included in its measurement model (EQ-5D-5L). The purposes of this study were (1) to assess the properties of the EQ-5D-5L in comparison with the standard EQ-5D-3L in a sample of young adults, (2) to foreground the importance of collecting qualitative data to confirm, validate or refine the EQ-5D questionnaire items and (3) to raise questions pertaining to the wording in these questionnaire items. Methods The data used came from a sample of respondents aged 30 or under (n = 624). They completed both versions of the EQ-5D, which were compared in terms of feasibility, level of inconsistency and ceiling effect. Agreement between the instruments was assessed using correlation coefficients and Bland-Altman plots. Known-groups validity of the EQ-5D-5L was also assessed using non-parametric tests. The discriminative properties were compared using receiver operating characteristic curves. Finally, four interviews were conducted for retrospective reports to elicit respondents’ understanding and perceptions of the format, instructions, items, and responses. Results Quantitative results show a ceiling effect reduction of 25.3 % and a high level agreement between both indices. Known-groups validity was confirmed for the EQ-5D-5L. Explorative interviews indicated ambiguity and low degree of certainty in regards to conceptualizing differences between levels moderate-slight across three dimensions. Conclusions The EQ-5D-5L performed better than the EQ-5D-3L. However, the explorative interviews demonstrated several limitations in the EQ-5D questionnaire wording and high context-dependent answers point to lack of illnesses’ experience amongst young adults.

Relationship between menopause and periodontal disease: a cross-sectional study in a Portuguese population

Article published under a “Creative Commons Attribution Noncommercial License”, enabling the unrestricted non-commercial use, distribution, and reproduction of the published article in any medium, provided that the original work is properly cited.

Avaliação do desempenho das entidades gestoras dos serviços de resíduos urbanos em Portugal

Dissertação de Mestrado, Gestão Empresarial, Faculdade de Economia, Universidade do Algarve, 2016

Plasma D-lactate Levels in Necrotizing Enterocolitis in Premature Infants

Background: D-Lactate is normally present in the blood of humans at nanomolar concentrations due to methylglyoxal metabolism; millimolar D-lactate concentrations can arise due to excess gastrointestinal microbial production. Objectives: To examine the levels of plasma D-lactate in the necrotizing enterocolitis in premature infants. Patients and Methods: 128 premature infants were divided into control (group I, n = 69), feeding intolerance (group II, n = 42) and NEC (group III, n = 27) groups. Plasma D-lactate levels were measured at the onset of feeding intolerance or NEC and at weeks 2-3 in control infants (group I) by ELISA. Data were analyzed using descriptive statistics, non-parametric tests and Student’s t-test. Results: In groups I, II, III, median birth weights were 1845.7 ± 267.5 g, 1913.1 ± 306.5 g, and 1898.4 ± 285.3 g, median gestational ages were 34.3 ± 1.7 weeks, 33.9 ± 2.2 weeks and 35.1 ± 2.6 weeks, ages of sampling were 12.3 ± 2.9 days, 14.6 ± 3.7 days and 15.1 ± 1.8 days, respectively. The differences of median birth weights, median gestational ages and ages of sampling were not statistically significant (P > 0.05). The plasma D-lactate levels in groups I, II, III were 3.6 ± 1.9 μg/mL, 12.7 ± 8.3 μg/mL, and 35.4 ± 29.1 μg/mL, respectively, group III had higher plasma D-lactate level than groups I, II, and the difference among these groups was significant (x2 = 21.6, P < 0.01). Conclusions: Plasma D-lactate significantly increased early in NEC. Plasma D-lactate levels were associated with extensive disease in NEC infants. Therefore, it could be used as a diagnosis indicator in the early stage of NEC.

Timing versus caloric goal in nutritional therapy for critically ill patients

Introduction: Enteral nutrition is an important therapy for severely critically ill patients. The timing and amount of energy have been highly debated. Objective: The aim of the present study was to directly compare the timing and the caloric targets in critically ill patients. Methods: Retrospective cohort study conducted at a single center, comparing timing and caloric goal for critically ill patients. Patients were stratified according to the start of nutritional therapy (24, 48, or more than 48 h) and the amount of energy delivered (target adequacy of previously calculated percentage in the first week). Statistical analysis was performed using parametric and non-parametric tests for independent samples and logistic regression. The results were expressed as mean ± standard deviation or incidence and percentage. Results and discussion: There were no differences in major clinical outcomes in relation to the achievement of percentage of caloric goal at the end of the first week of the study. The beginning of caloric intake on the first day of hospitalization was associated with reduced mortality in the intensive care unit, but not with hospital mortality. The strategy of an early and limited amount of calories seems to be associated with a better outcome. Prospective studies evaluating and comparing these strategies are recommended.

Nurses knowledge of health effects of dietary fibre

Introduction: Nurses accompany patients throughout their health care to prevent and treat disease, so their knowledge about diet and dietary fibre is key to successful diet therapy, which is an essential part of a patient's non-pharmacological treatment. It is known from the literature that a high proportion of nurses have limited knowledge about diet therapy and about sources of soluble fibre and other foods that can prevent or treat certain diseases. Given the position of nurses as key providers of dietary guidance, and given the health benefits of dietary fibre, we wanted to assess the level of fibre-related knowledge among nurses in Croatia. Material and Methods: Cross-sectional study based on data collected between October 2014 and March 2015 using a survey developed by the CI&DETS Polytechnic Institute in Viseu, Portugal. The survey contains questions about demographic characteristics as well as about knowledge of sources of dietary fibre, recommended daily intake and effects of fibre intake on particular diseases. The study included a total of 369 nurses from two health institutions and one nursing school from Croatia older than 18 years. Differences in knowledge were assessed for significance using the non-parametric Mann-Whitney U test. Possible associations among variables were explored using Spearman's rank correlation. For all statistical analyses, the threshold of significance was defined as P<0.05. Results: The level of knowledge among nurses in Croatia about dietary fibre varied from «undecided» to «partial knowledge». The median for level of knowledge ranged from 3 to 4 with low variability ranging from 0.11 to 0.33. Average levels of knowledge in percentages varied from 57.6% to 82.1%. Nurses with higher education levels showed significantly higher knowledge levels about the influence of dietary fiber intake on the risk of certain diseases (p = 0.007), constipation (p = 0.016), bowel cancer (p = 0.005) and breast cancer (p = 0.039). Conclusion: The level of nurses’ knowledge about dietary fiber is suboptimal. This indicates the need to strengthen nurse education in the areas of diet and diet therapy. Increase the level of knowledge of nurses about nutrition can positively influence the quality of care.

Hilbert Space Valued Signal Plus Noise Models: Analysis of Structural Breaks under High Dimensionality and Temporal Dependence

This thesis is concerned with change point analysis for time series, i.e. with detection of structural breaks in time-ordered, random data. This long-standing research field regained popularity over the last few years and is still undergoing, as statistical analysis in general, a transformation to high-dimensional problems. We focus on the fundamental »change in the mean« problem and provide extensions of the classical non-parametric Darling-Erdős-type cumulative sum (CUSUM) testing and estimation theory within highdimensional Hilbert space settings. In the first part we contribute to (long run) principal component based testing methods for Hilbert space valued time series under a rather broad (abrupt, epidemic, gradual, multiple) change setting and under dependence. For the dependence structure we consider either traditional m-dependence assumptions or more recently developed m-approximability conditions which cover, e.g., MA, AR and ARCH models. We derive Gumbel and Brownian bridge type approximations of the distribution of the test statistic under the null hypothesis of no change and consistency conditions under the alternative. A new formulation of the test statistic using projections on subspaces allows us to simplify the standard proof techniques and to weaken common assumptions on the covariance structure. Furthermore, we propose to adjust the principal components by an implicit estimation of a (possible) change direction. This approach adds flexibility to projection based methods, weakens typical technical conditions and provides better consistency properties under the alternative. In the second part we contribute to estimation methods for common changes in the means of panels of Hilbert space valued time series. We analyze weighted CUSUM estimates within a recently proposed »high-dimensional low sample size (HDLSS)« framework, where the sample size is fixed but the number of panels increases. We derive sharp conditions on »pointwise asymptotic accuracy« or »uniform asymptotic accuracy« of those estimates in terms of the weighting function. Particularly, we prove that a covariance-based correction of Darling-Erdős-type CUSUM estimates is required to guarantee uniform asymptotic accuracy under moderate dependence conditions within panels and that these conditions are fulfilled, e.g., by any MA(1) time series. As a counterexample we show that for AR(1) time series, close to the non-stationary case, the dependence is too strong and uniform asymptotic accuracy cannot be ensured. Finally, we conduct simulations to demonstrate that our results are practically applicable and that our methodological suggestions are advantageous.