Search

In the literature, there are two probabilistic models of bivariate Poisson : the model according to Holgate and the model according to Berkhout and Plug. These two models express themselves by their probability mass function. The model of Holgate puts in evidence a strictly positive correlation, which is not always realistic. To remedy this problem, Berkhout and Plug proposed a bivariate Poisson distribution accepting the correlation as well negative, equal to zero, that positive. In this paper, we show that these models are nearly everywhere asymptotically equal. From this survey that the o- divergence converges toward zero, both models are therefore nearly everywhere equal. Also, the model of Holgate converges toward the one of Berkhout and Plug. Some graphs will be presented for illustrating this comparison. Keywords: Correlation structure; distribution convergence

The Batch Markov Modulated Poisson Process (BMMPP) is a subclass of the versatile Batch Markovian Arrival process (BMAP) which have been widely used for the modeling of dependent and correlated simultaneous events (as arrivals, failures or risk events, real-time multimedia communications). Both theoretical and applied aspects are examined in this paper. On one hand, the identifiability of the stationary BMMPP2(K) is proven, where K is the maximum batch size. This is a powerful result when inferential tasks related to real data sets are carried out. On the other hand, some findings concerning the correlation and autocorrelation structures are provided. The first and second authors acknowledge financial support from the Spanish Ministry of Economy and Competitiveness, research project ECO2015-66593-P. The Third author acknowledge financial support from the Spanish Ministry of Economy and Competitiveness, research project MTM2015-65915-R; and also from Junta de Andalucía, and BBVA Fundation, research project P11- FQM-7603 and FQM-329

In distributed and decentralized state estimation systems, fusion methods are employed to systematically combine multiple estimates of the state into a single, more accurate estimate. An often encountered problem in the fusion process relates to unknown common information that is shared by the estimates to be fused and is responsible for correlations. If the correlation structure is unknown to the fusion method, conservative strategies are typically pursued. As such, the parameterization introduced by the ellipsoidal intersection method has been a novel approach to describe unknown correlations, though suitable values for these parameters with proven consistency have not been identified yet. In this article, an extension of ellipsoidal intersection is proposed that guarantees consistent fusion results in the presence of unknown common information. The bound used by the novel approach corresponds to computing an outer ellipsoidal bound on the intersection of inverse covariance ellipsoids. As a major advantage of this inverse covariance intersection method, fusion results prove to be more accurate than those provided by the well-known covariance intersection method. © 2017 Elsevier Ltd

In distributed and decentralized state estimation systems, fusion methods are employed to systematically combine multiple estimates of the state into a single, more accurate estimate. An often encountered problem in the fusion process relates to unknown common information that is shared by the estimates to be fused and is responsible for correlations. If the correlation structure is unknown to the fusion method, conservative strategies are typically pursued. As such, the parameterization introduced by the ellipsoidal intersection method has been a novel approach to describe unknown correlations, though suitable values for these parameters with proven consistency have not been identified yet. In this article, an extension of ellipsoidal intersection is proposed that guarantees consistent fusion results in the presence of unknown common information. The bound used by the novel approach corresponds to computing an outer ellipsoidal bound on the intersection of inverse covariance ellipsoids. As a major advantage of this inverse covariance intersection method, fusion results prove to be more accurate than those provided by the well-known covariance intersection method. © 2017 Elsevier Ltd

For decentralized fusion problems, ellipsoidal intersection has been proposed as an efficient fusion rule that provides less conservative results as compared to the well-know covariance intersection method. Ellipsoidal intersection relies on the computation of a common estimate that is shared by the estimates to be fused. In this paper, an algebraic reformulation of ellipsoidal intersection is discussed that circumvents the computation of the common estimate. It is shown that ellipsoidal intersection corresponds to an internal ellipsoidal approximation of the intersection of covariance ellipsoids. An interesting result is that ellipsoidal intersection can be computed with the aid of the Bar-Shalom/Campo fusion formulae. This is achieved by assuming a specific correlation structure between the estimates to be fused.

The Kalman filter correlated noises can be considered as an alternative to conventional filtering solutions, because it improves the positioning accuracy. The applications of this new filter would be very interesting for the coupling of measures in the following areas, photogrammetry, aerial remote sensing, bathymetry and «mobile mapping». The objective of this work graduation, is to assess the Kalman filter performance correlated noisesin an auscultation experience. In order to demonstrate the power of this filter, it must estimate by the maximum likelihood, one hand the variance of the evolution model and the other hand, the correlation models (same noises correlation structure for the evolution and observation models) setting beforehand a correlation structure to said models.; Le filtre de Kalman à bruits corrélés peut être considéré comme une alternative aux solutions de filtrage classique, car il améliore la précision du positionnement. Les applications de ce nouveau filtre seraient très intéressantes pour le couplage de mesures dans les domaines suivants, photogrammétrie, télédétection aérienne, bathymétrie et « mobile mapping ». L’objectif visé par ce travail de fin d’études, est d’évaluer les performances du filtre de Kalman à bruits corrélés dans une expérience d’auscultation. Dans l’optique de mettre en évidence la puissance de ce filtre, il y a lieu d’estimer par le maximum de vraisemblance, d’une part la variance du modèle d’évolution et d’autre part la corrélation des modèles ( même corrélation de bruits, pour les modèles d’évolution et d’observation) en fixant au préalable une structure de corrélation aux dits modèles.

Feature selection is an important prerequisite of any pattern recognition, machine learning or data mining problem. A lot of algorithms for feature subset selection have been developed so far for reduction of dimensionality of the data set in order to achieve high recognition accuracy with low computational cost. However, some methods or algorithms work well for some of the data sets and perform poorly on others. For any particular data set, it is difficult to find out the most suitable algorithm without some random trial and error process. It seems that the characteristics of the data set might have some effect on the algorithm for feature selection. In this work, the data set characteristics is studied for recommendation of appropriate feature selection algorithm to be used for a particular data set. A new proposal in terms of intra attribute relationship and a measure MVS (multivariate score) has been introduced to quantify and group different data sets on the basis of the data set correlation structure into several categories. The measure is used to group 63 publicly available bench mark data set according to their characteristics. The performance of different feature selection algorithms on different groups of data are then studied by simulation experiments to verify the relationship o f data set characteristics and the feature selection algorithm. The effect of some other data set characteristics has also been studied. Finally a framework of recommendation regarding the choice of proper feature selection algorithm has been indicated.

A seizure prediction algorithm is proposed that combines novel multivariate EEG features with patient-specific machine learning. The algorithm computes the eigenspectra of space–delay correlation and covariance matrices from 15‐s blocks of EEG data at multiple delay scales. The principal components of these features are used to classify the patient's preictal or interictal state. This is done using a support vector machine (SVM), whose outputs are averaged using a running 15-minute window to obtain a final prediction score. The algorithm was tested on 19 of 21 patients in the Freiburg EEG data set who had three or more seizures, predicting 71 of 83 seizures, with 15 false predictions and 13.8h in seizure warning during 448.3h of interictal data. The proposed algorithm scales with the number of available EEG signals by discovering the variations in correlation structure among any given set of signals that correlate with seizure risk.

Background: In age-related macular degeneration (ARMD) trials, the FDA-approved endpoint is the loss (or gain) of at least three lines of vision as compared to baseline. The use of such a response endpoint entails a potentially severe loss of information. A more efficient strategy could be obtained by using longitudinal measures of the change in visual acuity. In this paper we investigate, by using data from two randomized clinical trials, the mean and variance–covariance structures of the longitudinal measurements of the change in visual acuity. Methods: Individual patient data were collected in 234 patients in a randomized trial comparing interferon-α with placebo and in 1181 patients in a randomized trial comparing three active doses of pegaptanib with sham. A linear model for longitudinal data was used to analyze the repeated measurements of the change in visual acuity. Results: For both trials, the data were adequately summarized by a model that assumed a quadratic trend for the mean change in visual acuity over time, a power variance function, and an antedependence correlation structure. The power variance function was remarkably similar for the two datasets and involved the square root of the measurement time. Conclusions: The similarity of the estimated variance functions and correlation structures for both datasets indicates that these aspects may be a genuine feature of the measurements of changes in visual acuity in patients with ARMD. The feature can be used in the planning and analysis of trials that use visual acuity as the clinical endpoint of interest. Copyright © 2010 John Wiley & Sons, Ltd.

Contrary to Schein’s theory of career anchors, which rests on the dominance of a single career anchor, the present study proposes an original career anchor structure that captures multiple dominant anchors. The analysis of data from a sample of 880 Quebec engineers supports this reconceptualization based on a circular model of career anchors. The new dynamics of career anchors shows that several anchors are complementary (e.g., creativity and challenge) while others are conflictual (e.g., challenge and security). In particular, the correlational analysis at the axial level indicates that the “self-enhancement” pole (managerial competence, identity) is negatively correlated with the “self-transcendence” (service/dedication to a cause, technical competence), whereas the pole “openness to change” (challenge, entrepreneurial creativity) is negatively correlated with the “conservation” pole (security, lifestyle). These findings can lead to more research in career management., Le modèle des ancres de carrière de Schein (1978) propose que les individus ne possèdent qu’une seule ancre dominante, celle-ci reflétant les habiletés, les valeurs et les besoins fondamentaux privilégiés par une personne au cours de sa carrière. Un tel phénomène de dominance, souvent opérationnalisé en identifiant l’ancre ayant obtenu le score le plus élevé chez un individu, est appelé « différenciation ». Récemment, plusieurs chercheurs ont remis en question la différenciation. Selon Martineau, Wils et Tremblay (2005), certains individus intériorisent en fait plusieurs ancres élevées, ce qui fait référence à « l’indifférenciation ».Contrairement au modèle de Schein, celui de Feldman et Bolino (1996) aide à comprendre l’indifférenciation parce qu’il repose sur une structuration des ancres organisée selon une dynamique octogonale. En effet, il existerait une proximité de certaines ancres dites « compatibles » (ancres connexes de l’octogone) ainsi qu’une opposition entre d’autres ancres dites « incompatibles » (ancres diamétralement opposées de l’octogone). L’idée de structuration des ancres selon une dynamique circulaire est théoriquement intéressante, car elle pourrait expliquer pourquoi un individu n’afficherait que quelques ancres dominantes données. Vu que la justification avancée par Feldman et Bolino pour articuler les ancres ne découle pas d’une preuve empirique probante, leur modèle a besoin d’être peaufiné d’un point de vue théorique en faisant appel à un autre modèle de structuration. Les ancres étant essentiellement des valeurs qui guident les décisions de carrière et, par le fait même, l’évolution de la carrière et du travail des individus, il est logique de mobiliser le modèle de structuration des valeurs de Schwartz. En particulier, un modèle de structuration circulaire des ancres émerge d’un parallèle entre les valeurs et les ancres de carrière.Plusieurs études empiriques ont montré soit des corrélations positives, soit des corrélations négatives entre des ancres de carrière. Sans un modèle de structuration circulaire des ancres, il devient difficile d’interpréter ces corrélations d’autant plus qu’elles proviennent de données n’ayant pas été normalisées sur une base individuelle comme le recommande Schwartz (1992). Aussi l’objectif de cette étude est-il de vérifier l’existence d’une telle structuration des ancres en mettant à jour la dynamique corrélationnelle entre les ancres. Pour ce faire, cette recherche a adopté la méthodologie développée par Schwartz. Les données, qui proviennent d’un échantillon de 880 ingénieurs québécois, sont soumises à des analyses multidimensionnelles (de type SSA « Smallest Space Analysis » de Guttman-Lingoes, disponible uniquement avec SYSTAT) ainsi qu’à des analyses corrélationnelles (SPSS) afin de vérifier six hypothèses.Les résultats des analyses multidimensionnelles sont représentés sur une carte composée de quatre quadrants. Un premier quadrant correspond au pôle « affirmation de soi ». Ce quadrant regroupe des items reliés aux domaines du pouvoir (richesse, autorité, pouvoir social et reconnaissance sociale) et d’accomplissement (réussite, influence). Ainsi, l’ancre de gestion et l’ancre d’identité font partie d’une même famille d’ancres managériales. Opposé à ce quadrant, se situe le domaine associé au pôle « dépassement de soi » composé des items reliés au domaine de la quête de savoir (ancre de compétence technique) et de la bienveillance (ancre service/dévouement à une cause). D’autre part, deux autres quadrants s’opposent. Un troisième quadrant rassemble l’item associé au domaine de la sécurité (ancre de sécurité) auquel s’ajoute les items ayant trait à la QVT, aux conditions de travail et au climat de travail (ancre style de vie); il s’agit donc du pôle « continuité ». Enfin, un quatrième quadrant, celui de l’ouverture au changement, englobe des items reliés à l’auto-orientation (ancre de créativité) et à la stimulation (ancre de défi) auxquels s’ajoutent les items reliés au changement d’emploi, au changement géographique et au changement de projet. De ces résultats il en découle que les quatre premières hypothèses reçoivent un fort appui empirique.Les deux dernières hypothèses de cette étude portent sur la dynamique corrélationnelle. À partir des items identifiés dans chacun des quadrants, quatre variables composites ont été calculées après avoir normalisé les données par individu. Ces quatre variables correspondent aux quatre pôles, à savoir l’affirmation de soi, le dépassement de soi, l’ouverture au changement et la continuité. L’analyse corrélationnelle au niveau axial indique que le pôle « affirmation de soi » est négativement corrélé au pôle de dépassement de soi (r = -0,44 significatif à p, Contrariamente a la teoría de anclaje de carreras de Schein, que se queda en el campo de un anclaje único de carrera, el presente estudio propone una estructura de anclaje de carrera original que incluye múltiples anclajes dominantes. El análisis de datos a partir de una muestra de 880 ingenieros de Quebec sostiene esta re-conceptualización basada en un modelo circular de anclajes de carrera. Las nuevas dinámicas de anclaje de carrera muestran que varios anclajes son complementarios (por ejemplo, creatividad y reto) mientras otros son conflictivos (por ejemplo, reto y seguridad). En particular, el análisis de correlación al nivel axial indica que el polo “auto-mejora” (dedicación a la causa, competencia técnica) está en correlación negativa con la “auto-trascendencia”, mientras el polo “apertura al cambio” (reto, creatividad empresarial) está en correlación negativa con el polo “conservador” (seguridad, estilo de vida). Estos resultados pueden llevar a otras investigaciones sobre la carrera de gestión.

We are concerned with statistical inference for 2 × C × K contingency tables in the context of genetic case-control association studies. Multivariate methods based on asymptotic Gaussianity of vectors of test statistics require information about the asymptotic correlation structure among these test statistics under the global null hypothesis. In the case of C=2, we show that for a wide variety of test statistics this asymptotic correlation structure is given by the standardized linkage disequilibrium matrix of the K loci under investigation. Three popular choices of test statistics are discussed for illustration. In the case of C=3, the standardized composite linkage disequilibrium matrix is the limiting correlation matrix of the K locus-specific Cochran-Armitage trend test statistics.

The spatial variability of stress fields resulting from polycrystalline aggregate calculations involving random grain geometry and crystal orientations is investigated. A periodogram-based method is proposed to identify the properties of homogeneous Gaussian random fields (power spectral density and related covariance structure). Based on a set of finite element polycrystalline aggregate calculations the properties of the maximal principal stress field are identified. Two cases are considered, using either a fixed or random grain geometry. The stability of the method w.r.t the number of samples and the load level (up to 3.5% macroscopic deformation) is investigated., Frontiers of Structural and Civil Engineering, 9 (2), ISSN:2095-2430, ISSN:2095-2449

The multifractional Brownian motion (MBM) processes are locally self-similar Gaussian processes. They extend the classical fractional Brownian motion processes B H = { B H ( t ) } t ∈ R by allowing their self-similarity parameter H ∈ ( 0 , 1 ) to depend on time. Two types of MBM processes were introduced independently by Peltier and Levy-Vehel [Multifractional Brownian motion: definition and preliminary results, Technical Report 2645, Institut National de Recherche en Informatique et an Automatique, INRIA, Le Chesnay, France, 1995] and Benassi, Jaffard, Roux [Elliptic Gaussian random processes, Rev. Mat. Iber. 13(1) (1997) 19–90] by using time-domain and frequency-domain integral representations of the fractional Brownian motion, respectively. Their correspondence was studied by Cohen [From self-similarity to local self-similarity: the estimation problem, in: M. Dekking, J.L. Vehel, E. Lutton, C. Tricot (Eds.), Fractals: Theory and Applications in Engineering, Springer, Berlin, 1999]. Contrary to what has been stated in the literature, we show that these two types of processes have different correlation structures when the function H ( t ) is non-constant. We focus on a class of MBM processes parameterized by ( a + , a - ) ∈ R 2 , which contains the previously introduced two types of processes as special cases. We establish the connection between their time- and frequency-domain integral representations and obtain explicit expressions for their covariances. We show, that there are non-constant functions H ( t ) for which the correlation structure of the MBM processes depends non-trivially on the value of ( a + , a - ) and hence, even for a given function H ( t ) , there are an infinite number of MBM processes with essentially different distributions.

We study the no-arbitrage theory of voluntary disclosure (Dye, J Account Res 23:123-145, 1985, and Ostaszewski and Gietzmann, Rev Quant Financ Account 31: 1-27, 2008), generalized to the setting of {Mathematical expression} firms, simultaneously and voluntarily, releasing at the interim-report date 'partial' information concerning their 'common operating conditions'. Each of the firms has, as in the Dye model, some (known) probability of observing a signal of their end of period performance, but here this signal includes noise determined by a firm-specific precision parameter. The co-dependency of the firms results entirely from their common operating conditions. Each firm has a disclosure cutoff, which is a best response to the cutoffs employed by the remaining firms. To characterize these equilibrium cutoffs explicitly, we introduce {Mathematical expression} new hypothetical firms, related to the corresponding actual firms, which are operationally independent, but are assigned refined precision parameters and amended means. This impounds all existing correlations arising from conditioning on the other potentially available sources of information. In the model the actual firms' equilibrium cutoffs are geometric weighted averages of these hypothetical firms. We uncover two countervailing effects. Firstly, there is a bandwagon effect, whereby the presence of other firms raises each individual cutoff relative to what it would have been in the absence of other firms. Secondly, there is an estimator-quality effect, whereby individual cutoffs are lowered, unless the individual precision is above average.

Wildlife population parameters, such as capture or detection probabilities, and density or population size, can be estimated from capture-recapture data. These estimates are of particular interest to ecologists and biologists who rely on ac- curate inferences for management and conservation of the population of interest. However, there are many challenges to researchers for making accurate inferences on population parameters. For instance, capture-recapture data can be considered as binary longitudinal observations since repeated measurements are collected on the same individuals across successive points in times, and these observations are often correlated over time. If these correlations are not taken into account when estimating capture probabilities, then parameter estimates will be biased, possibly producing misleading results. Also, an estimator of population size is generally biased under the presence of heterogeneity in capture probabilities. The use of covariates (or auxiliary variables), when available, has been proposed as an alternative way to cope with the problem of heterogeneous capture probabilities. In this dissertation, we are interested in tackling these two main problems, (i) when capture probabilities are dependent among capture occasions in closed population capture-recapture models, and (ii) when capture probabilities are heterogeneous among individuals. Hence, the capture-recapture literature can be improved, if we could propose an approach to jointly account for these problems. In summary, this dissertation proposes: (i) a generalized estimating equations (GEE) approach to model possible effects in capture-recapture closed population studies due to correlation over time and individual heterogeneity; (ii) the corresponding estimating equations for each closed population capture-recapture model; (iii) a comprehensive analysis on various real capture-recapture data sets using classical, GEE and generalized linear mixed models (GLMM); (iv) an evaluation of the effect of ac- counting for correlation structures on capture-recapture model selection based on the ‘Quasi-likelihood Information Criterion (QIC)’; (v) a comparison of the performance of population size estimators using GEE and GLMM approaches in the analysis of capture-recapture data. The performance of these approaches is evaluated by Monte Carlo (MC) simulation studies resembling real capture-recapture data. The proposed GEE approach provides a useful inference procedure for estimating population parameters, particularly when a large proportion of individuals are captured. For a low capture proportion, it is difficult to obtain reliable estimates for all approaches, but the GEE approach outperforms the other methods. Simulation results show that quasi-likelihood GEE provide lower standard error than partial likelihood based on generalized linear modelling (GLM) and GLMM approaches. The estimated population sizes vary on the nature of the existing correlation among capture occasions; RESUMO: Parâmetros populacionais em espécies de vida selvagens, como probabilidade captura ou deteção, e abundância ou densidade da população, podem ser estimados a partir de dados de captura-recaptura. Estas estimativas são de particular interesse para ecologistas e biólogos que dependem de inferências precisas a gestão e conservação das populações. No entanto, há muitos desafios par investigadores fazer inferências precisas de parâmetros populacionais. Por exemplo, os dados de captura-recaptura podem ser considerados como observa longitudinais binárias uma vez que são medições repetidas coletadas nos mesmos indivíduos em pontos sucessivos no tempo, e muitas vezes correlacionadas. Essas correlações não são levadas em conta ao estimar as probabilidades de tura, as estimativas dos parâmetros serão tendenciosas e possivelmente produz resultados enganosos. Também, um estimador do tamanho de uma população geralmente enviesado na presença de heterogeneidade das probabilidades de captura. A utilização de co-variáveis (ou variáveis auxiliares), quando disponível tem sido proposta como uma forma de lidar com o problema de probabilidade captura heterogéneas. Nesta dissertação, estamos interessados em abordar problemas principais em mode1os de captura-recapturar para população fecha (i) quando as probabilidades de captura são dependentes entre ocasiões de captura e (ii) quando as probabilidades de captura são heterogéneas entre os indivíduos Assim, a literatura de captura-recaptura pode ser melhorada, se pudéssemos por uma abordagem conjunta para estes problemas. Em resumo, nesta dissertação propõe-se: (i) uma abordagem de estimação de equações generalizadas (GEE) para modelar possíveis efeitos de correlação temporal e heterogeneidade individual nas probabilidades de captura; (ii) as correspondentes equações de estimação generalizadas para cada modelo de captura-recaptura em população fechadas; (iii) uma análise sobre vários conjuntos de dados reais de captura-recaptura usando a abordagem clássica, GEE e modelos linear generalizados misto (GLMM); (iv) uma avaliação do efeito das estruturas de correlação na seleção de modelos de captura-recaptura com base no ‘critério de informação da Quasi-verossimilhança (QIC); (v) uma comparação da performance das estimativas do tamanho da população usando GEE e GLMM. O desempenho destas abordagens ´e avaliado usando simulações Monte Carlo (MC) que se assemelham a dados de captura- recapture reais. A abordagem GEE proposto ´e um procedimento de inferência útil para estimar parâmetros populacionais, especialmente quando uma grande proporção de indivíduos ´e capturada. Para uma proporção baixa de capturas, ´e difícil obter estimativas fiáveis para todas as abordagens aplicadas, mas GEE supera os outros métodos. Os resultados das simulações mostram que o método da quase-verossimilhança do GEE fornece estimativas do erro padrão menor do que o método da verossimilhança parcial dos modelos lineares generalizados (GLM) e GLMM. As estimativas do tamanho da população variam de acordo com a natureza da correlação existente entre as ocasiões de captura.

Great progress has been made towards understanding the properties of single neurons, yet the principles underlying interactions between neurons remain poorly understood. Given that connectivity in the neocortex is locally dense through both horizontal and vertical con-nections, it is of particular importance to characterize the activity structure of local popula-tions of neurons arranged in three dimensions. However, techniques for simultaneously measuring microcircuit activity are lacking. We developed an in vivo 3D high-speed, ran-dom-access two-photon microscope that is capable of simultaneous 3D motion tracking. This allows imaging from hundreds of neurons at several hundred Hz, while monitoring tissue movement. Given that motion will induce common artifacts across the population, accurate motion tracking is absolutely necessary for studying population activity with ran-dom-access based imaging methods. We demonstrate the potential of this imaging tech-nique by measuring the correlation structure of large populations of nearby neurons in the mouse visual cortex, and find that the microcircuit correlation structure is stimulus-dependent. Three-dimensional random access multiphoton imaging with concurrent mo-tion tracking provides a novel powerful method to characterize the microcircuit activity in vivo.

Financiadora de Estudos e Projetos In this dissertation we propose new models for survival with cure fraction to describe the biological mechanism of the event of interest (cancer) in studies of carcinogenesis in the presence of competing causes latent independent or correlated. The formulation of new models is based on stochastic modeling of the occurrence of tumors through three stages: initiation of a tumor not detectable, promotion and progression of the tumor to a detectable cancer. These models allow a simple pattern of the dynamics of tumor growth, and incorporate into the analysis features of the stage of tumor progression that is not possible in most survival models with cure fraction commonly used. For the proposed models, the inferential process was discussed in terms of classical and Bayesian point of view. Simulations studies were conducted in order to analyze the asymptotical properties of the classical estimation procedure. Real data applications demonstrate of use of the models. Neste trabalho propomos modelos de sobrevivência com fração de cura para descrever o mecanismo biológico da ocorrência do evento de interesse (câncer) em estudos da carcinogênese na presença de causas competitivas latentes independentes ou correlacionadas. A formulação dos novos modelos é baseada na modelagem estocástica da ocorrência dos tumores através de três estágios: iniciação de um tumor não detectável, promoção e a progressão do tumor até um câncer detectável. Estes modelos permitem um padrão simples da dinâmica de crescimento do tumor, além de incorporarem características do estágio de progressão do tumor, que não é possível na maioria dos modelos de sobrevivência com fração de cura comumente utilizados. Para os modelos propostos, discutimos o processo inferencial do ponto de vista clássico e bayesiano. Estudos de simulações foram feitos com o objetivo de analisar as propriedades assintóticas do processo de estimação clássico. Aplicações a conjuntos de dados reais mostraram a aplicabilidade dos modelos.

International audience; The spatial variability of stress fields resulting from polycrystalline aggregate calculations involving random grain geometry and crystal orientations is investigated. A periodogram-based method is proposed to identify the properties of homogeneous Gaussian random fields (power spectral density and related covariance structure). Based on a set of finite element polycrystalline aggregate calculations the properties of the maximal principal stress field are identified. Two cases are considered, using either a fixed or random grain geometry. The stability of the method w.r.t the number of samples and the load level (up to 3.5 \% macroscopic deformation) is investigated.

Using weekly returns of S&P 500 constituents, we study the time-varying correlation structure during the period of 2006 to mid-2011. Contrary to most of the previous correlation studies of many assets, we do not use rolling correlations but the DCC MV-GARCH model with the MacGyver strategy proposed by Engle (2009). We find empirical evidence that the correlation structure tends to change significantly during the periods of high volatility and market downturns.

In longitudinal studies subjects are measured for one or more response variable, over time. Although the underlying evolution of such response variables is continuous in time, in practice the measurements are observed at discrete time points. In longitudinal clinical trials it is also common to observe relevant events, generating time-to-event data. If both types of data are available, we might be interested in the association between the two processes, longitudinal and time-to-event. Commonly, when death is considered the event, the observation sequence of longitudinal measurements is terminated by the event process. When the two observed processes are related, the analysis of the data set should be suited to the specific objectives. We distinguish three situations: if the interest is to analyse the longitudinal outcome response variable with drop-out at the time-to-event; to analyse time-to-event, whilst exploiting correlation with a noisy version of a time-varying risk factor; or to analyse the relationship between the two processes. Joint models assume a full distribution for the joint distribution of longitudinal and time-to-event processes, which includes a description of the relation between the two processes., This work has been supported by the grant number SFRH/BD/10266/2001 from FCT-Portugal, that supported the PhD of the author.

The use of improved covariance matrix estimators as an alternative to the sample estimator is considered an important approach for enhancing portfolio optimization. Here we empirically compare the performance of 9 improved covariance estimation procedures by using daily returns of 90 highly capitalized US stocks for the period 1997-2007. We find that the usefulness of covariance matrix estimators strongly depends on the ratio between estimation period T and number of stocks N, on the presence or absence of short selling, and on the performance metric considered. When short selling is allowed, several estimation methods achieve a realized risk that is significantly smaller than the one obtained with the sample covariance method. This is particularly true when T/N is close to one. Moreover many estimators reduce the fraction of negative portfolio weights, while little improvement is achieved in the degree of diversification. On the contrary when short selling is not allowed and T>N, the considered methods are unable to outperform the sample covariance in terms of realized risk but can give much more diversified portfolios than the one obtained with the sample covariance. When T, Comment: 30 pages

This paper proposes a detailed analysis of the covariance structure of the Cross-Nested Logit model, widely applied thanks to its closed-form and to the expected flexibility of the underlying covariance matrix structure. In more detail, through the numerical computation of actual CNL covariances, some relevant aspects are analysed and addressed. Firstly, the degree of approximation inherent Papola’s conjecture [Papola, A., 2004. Some developments of the Cross-Nested Logit model. Transportation Research Part B 38 (14), 833–851] on CNL covariances is explored. Then, it is shown that the CNL model is not generally able to cover the whole domain of the feasible homoskedastic covariance matrices, and that the degree of coverage depends on the adopted CNL nesting structure. In that respect, the issue of finding the most flexible CNL structure is addressed from a theoretical and numerical standpoint, leading to the result that maximum flexibility is reached for a “full” nesting structure (i.e. each alternative belongs to all the groups): a rule-of-thumb is also established for the choice of the number of groups. Moreover, it is shown that, when a covariance matrix is reproducible, there are generally infinite CNL specifications – leading to different choice probabilities – corresponding to that matrix. A direct consequence is that the issue of finding a CNL model specification able to reproduce a given known correlation matrix (relevant in some contexts, e.g. route choice modelling) can be not so practically relevant since choice probabilities are not in a one-to-one correspondence with covariances.

It is well known that correlations in microarray data represent a serious nuisance deteriorating the performance of gene selection procedures. This paper is intended to demonstrate that the correlation structure of microarray data provides a rich source of useful information. We discuss distinct correlation substructures revealed in microarray gene expression data by an appropriate ordering of genes. These substructures include stochastic proportionality of expression signals in a large percentage of all gene pairs, negative correlations hidden in ordered gene triples, and a long sequence of weakly dependent random variables associated with ordered pairs of genes. The reported striking regularities are of general biological interest and they also have far-reaching implications for theory and practice of statistical methods of microarray data analysis. We illustrate the latter point with a method for testing differential expression of nonoverlapping gene pairs. While designed for testing a different null hypothesis, this method provides an order of magnitude more accurate control of type 1 error rate compared to conventional methods of individual gene expression profiling. In addition, this method is robust to the technical noise. Quantitative inference of the correlation structure has the potential to extend the analysis of microarray data far beyond currently practiced methods., Comment: Published in at http://dx.doi.org/10.1214/07-AOAS120 the Annals of Applied Statistics (http://www.imstat.org/aoas/) by the Institute of Mathematical Statistics (http://www.imstat.org)

In this paper we propose a jump-diffusion Libor model with jumps in a high-dimensional space and test a stable non-parametric calibration algorithm which takes into account a given local covariance structure. The algorithm returns smooth and simply structured Lévy densities, and penalizes the deviation from the Libor market model. In practice, the procedure is FFT based, thus fast, easy to implement, and yields good results, particularly in view of the ill-posedness of the underlying inverse problem.

In this paper we propose a jump-diffusion Libor model with jumps in a high-dimensional space and test a stable non-parametric calibration algorithm which takes into account a given local covariance structure. The algorithm returns smooth and simply structured Lévy densities, and penalizes the deviation from the Libor market model. In practice, the procedure is FFT based, thus fast, easy to implement, and yields good results, particularly in view of the ill-posedness of the underlying inverse problem.

Gaussian traffic models are capable of representing a broad variety of correlation structures, ranging from short-range dependent (e.g., Ornstein-Uhlenbeck type) to long-range dependent (e.g., fractional Brownian motion, with Hurst parameter $H$ exceeding $\frac{1}{2}$). This note focuses on queues fed by a large number ($n$) of Gaussian sources, emptied at constant service rate $nc.$ In particular, we consider the probability of exceeding buffer level $nb,$ as a function of $b.$ This probability decaying (asymptotically) exponentially in $n,$ the essential information is contained in the exponential decay rate $I(b).$ The main result of this note describes the duality relation between the shape of $I(.)$ and the correlation structure. More specifically, it is shown that the curve $I(.)$ is convex at some buffer size $b$ if and only if there are negative correlations on the time scale at which the overflow takes place.

Simulating and understanding traffic dynamics in large networks are difficult and challenging due to the complexity of such networks and the limitations inherent in simulation modeling. Typically, simulation models used to study traffic dynamics include substantial detail representing protocol mechanisms across several layers of functionality. Such models must be restricted in space and time in order to be computationally tractable. We propose an alternative simulation approach that uses homogeneous modeling with an increased level of abstraction, in order to explore networks at larger space-time scales than otherwise feasible and to develop intuition and insight about the space-time dynamics of large networks. To illustrate the utility of our approach, we examine some current understandings of the timescale dynamics of network traffic, and we discuss some speculative results obtained with homogeneous modeling. Using a wavelet-based technique, we show correlation structures, and changes in correlation structures, of network traffic under variations in traffic sources, transport mechanisms, and network structure. Our simulation results justify further investigation of our approach, which might benefit from cross-verifications against more detailed simulation models.