FGV EMAp - Artigos


Submissões Recentes

Agora exibindo 1 - 20 de 30
  • Item
    On the normalized power prior
    (2020-04) Carvalho, Luiz Max de; Ibrahim, Joseph G.
    The power prior is a popular tool for constructing informative prior distributions based on historical data. The method consists of raising the likelihood to a discounting factor in order to control the amount of information borrowed from the historical data. It is customary to perform a sensitivity analysis reporting results for a range of values of the discounting factor. However, one often wishes to assign it a prior distribution and estimate it jointly with the parameters, which in turn necessitates the computation of a normalising constant. In this paper we are concerned with how to recycle computations from a sensitivity analysis in order to approximately sample from joint posterior of the parameters and the discounting factor. We first show a few important properties of the normalising constant and then use these results to motivate a bisection-type algorithm for computing it on a fixed budget of evaluations. We give a large array of illustrations and discuss cases where the normalising constant is known in closed-form and where it is not. We show that the proposed method produces approximate posteriors that are very close to the exact distributions when those are available and also produces posteriors that cover the data-generating parameters with higher probability in the intractable case. Our results show that proper inclusion the normalising constant is crucial to the correct quantification of uncertainty and that the proposed method is an accurate and easy to implement technique to include this normalisation, being applicable to a large class of models.
  • Item
    Challenges of evaluating and modelling vaccination in emerging infectious diseases
    (2021) Madewell, Zachary J.; Dean, Natalie E.; Berlin, Jesse A.; Coplan, Paul M.; Davis, Kourtney J.; Struchinerf, Claudio J.; Halloran, M. Elizabeth
    Outbreaks of emerging pathogens pose unique methodological and practical challenges for the design, implementation, and evaluation of vaccine efficacy trials. Lessons learned from COVID-19 highlight the need for innovative and flexible study design and application to quickly identify promising candidate vaccines. Trial design strategies should be tailored to the dynamics of the specific pathogen, location of the outbreak, and vaccine prototypes, within the regional socioeconomic constraints. Mathematical and statistical models can assist investigators in designing infectious disease clinical trials. We introduce key challenges for planning, evaluating, and modelling vaccine efficacy trials for emerging pathogens.
  • Item
    Prevalence and predictors of anti-SARS-CoV-2 serology in a highly vulnerable population of Rio de Janeiro: a population-based serosurvey
    (2022) Coelho, Lara E.; Luz, Paula M.; Pires, Débora C.; Jalil, Emilia M.; Perazzo, Hugo; Torres, Thiago S.; Cardoso, Sandra W.; Peixoto, Eduardo M.; Nazer, Sandro; Massad, Eduardo; Silveira, Mariângela F.; Barros, Fernando C.; Vasconcelos, Ana T. R.; Costa, Carlos A. M.; Amancio, Rodrigo T.; Villela, Daniel A. M.; Pereira, Tiago; Goedert, Guilherme T.; Santos, Cleber V. B. D.; Rodrigues, Nadia C. P.; Grinsztejn, Beatriz; Veloso, Valdiléa G.; Struchiner, Claudio J.
    COVID-19 serosurveys allow for the monitoring of the level of SARS-CoV-2 transmission and support data-driven decisions. We estimated the seroprevalence of anti-SARS-CoV-2 antibodies in a large favela complex in Rio de Janeiro, Brazil. Methods: A population-based panel study was conducted in Complexo de Manguinhos (16 favelas) with a probabilistic sampling of participants aged ≥1 year who were randomly selected from a census of individuals registered in primary health care clinics that serve the area. Participants answered a structured interview and provided blood samples for serology. Multilevel regression models (with random intercepts to account for participants' favela of residence) were used to assess factors associated with having anti-S IgG antibodies. Secondary analyses estimated seroprevalence using an additional anti-N IgG assay. Findings: 4,033 participants were included (from Sep/2020 to Feb/2021, 22 epidemic weeks), the median age was 39·8 years (IQR:21·8-57·7), 61% were female, 41% were mixed-race (Pardo) and 23% Black. Overall prevalence was 49·0% (95%CI:46·8%-51·2%) which varied across favelas (from 68·3% to 31·4%). Lower prevalence estimates were found when using the anti-N IgG assay. Odds of having anti-S IgG antibodies were highest for young adults, and those reporting larger household size, poor adherence to social distancing and use of public transportation. Interpretation: We found a significantly higher prevalence of anti-S IgG antibodies than initially anticipated. Disparities in estimates obtained using different serological assays highlight the need for cautious interpretation of serosurveys estimates given the heterogeneity of exposure in communities, loss of immunological biomarkers, serological antigen target, and variant-specific test affinity.
  • Item
    Incremental constraint projection methods for monotone stochastic variational inequalities
    (2017) Iusem, Alfredo Noel; Jofre, Alejandro; Thompson, Philip A.
    We consider stochastic variational inequalities with monotone operators defined as the expected value of a random operator. We assume the feasible set is the intersection of a large family of convex sets. We propose a method that combines stochastic approximation with incremental constraint projections meaning that at each iteration, a step similar to some variant of a deterministic projection method is taken after the random operator is sampled and a component of the intersection defining the feasible set is chosen at random. Such sequential scheme is well suited for applications involving large data sets, online optimization and distributed learning. First, we assume that the variational inequality is weak-sharp. We provide asymptotic convergene, feasibility rate of O(1/k) in terms of the mean squared distance to the feasible set and solvability rate of O(1/√k) (up to first order logarithmic terms) in terms of the mean distance to the solution set for a bounded or unbounded feasible set. Then, we assume just monotonicity of the operator and introduce an explicit iterative Tykhonov regularization to the method. We consider Cartesian variational inequalities so as to encompass the distributed solution of stochastic Nash games or multi-agent optimization problems under a limited coordination.
  • Item
    Optimal control of PDES in a complex space setting; application to the schrodinger equation
    (2016) Soledad Aronna, Maria; Bonnans, Frederic; Kroner, Axel
    In this paper we discuss optimality conditions for abstract optimization problems over complex spaces. We then apply these results to optimal control problems with a semigroup structure. As an application we detail the case when the state equation is the Schr¨odinger one, with pointwise constraints on the “bilinear” control. We derive first and second order optimality conditions and address in particular the case that the control enters the state equation and cost function linearly
  • Item
    A higher-order maximum principle for impulsive optimal control problems
    (2019) Soledad Aronna, Maria; Motta, Monica; Rampazzo, Franco
    We consider a nonlinear system, affine with respect to an unbounded control u which is allowed to range in a closed cone. To this system we associate a Bolza type minimum problem, with a Lagrangian having sublinear growth with respect to u. This lack of coercivity gives the problem an impulsive character, meaning that minimizing sequences of trajectories happen to converge towards discontinuous paths. As is known,a distributional approach does not make sense in such a nonlinear setting, where, instead, a suitable embedding in the graph-space is needed. We provide higher order necessary optimality conditions for properly defined impulsive minima, in the form of equalities and inequalities involving iterated Lie brackets of the dynamical vector fields. These conditions are derived under very weak regularity assumptions and without any constant rank conditions.
  • Item
    Non-asymptotic con dence bounds for the optimal value of a stochastic program
    (2016) Guigues, Vincent Gérard Yannick; Juditsky, Anatoli; Nemirovski, Arkadi Semenovich
    We discuss a general approach to building non-asymptotic con dence bounds for stochastic optimization problems. Our principal contribution is the observation that a Sample Average Approximation of a problem supplies upper and lower bounds for the optimal value of the problem which are essentially better than the quality of the corresponding optimal solutions. At the same time, such bounds are more reliable than \standard" con dence bounds obtained through the asymptotic approach. We also discuss bounding the optimal value of MinMax Stochastic Optimization and stochastically constrained problems. We conclude with a simulation study illustrating the numerical behavior of the proposed bounds.
  • Item
    Inexact cuts in stochastic dual dynamic programming
    (2019) Guigues, Vincent Gérard Yannick
    We introduce an extension of Stochastic Dual Dynamic Programming (SDDP) to solve stochastic convex dynamic programming equations. This extension applies when some or all primal and dual subproblems to be solved along the forward and backward passes of the method are solved with bounded errors (inexactly). This inexact variant of SDDP is described both for linear problems (the corresponding variant being denoted by ISDDP-LP) and nonlinear problems (the corresponding variant being denoted by ISDDP-NLP). We prove convergence theorems for ISDDP-LP and ISDDP-NLP both for bounded and asymptotically vanishing errors. Finally, we present the results of numerical experiments comparing SDDP and ISDDP-LP on a portfolio problem with direct transaction costs modelled as a multistage stochastic linear optimization problem. On these experiments, ISDDP-LP allows us to obtain a good policy faster than SDDP.
  • Item
    Uncertain spaces, uncertain places. Dealing with geographic information in digital humanities: the example of a language legacy dataset
    (GI_Forum, 2020) Dorn, Amelie; Souza, Renato Rocha; Piringer, Barbara; Wandl-Vogt, Eveline
    In addition to their purely linguistic content, legacy language collections often contain other information, such as geographical and spatial details, e.g. locations, regions and municipalities. Such information may offer valuable insights into the linguistic landscape, but it may also pose challenges when some aspects remain ambiguous. This paper outlines and discusses various known and unknown uncertainties of spatial aspects contained in a nonstandard German language legacy dataset (DBÖ) that has undergone several stages of data conversion since the early nineties. The authors introduce and discuss their taxonomy of uncertainties, exemplified by applying it to the spatial information contained in the DBÖ, the origins of which date back one hundred years. Finally, the authors discuss how the uncertainties found in the dataset affect Digital Humanities practice more widely.
  • Item
    Endogenous asymmetric money illusion
    (Elsevier, 2019) Duarte, Diogo; Saporito, Yuri Fahham
    We show that when investors suffer from endogenous asymmetric money illusion, the usual proportionality between money supply and nominal prices commonly present in frictionless economies is eliminated. This drives changes in the money supply to cause real price fluctuations. Nevertheless, the combined effect on the real state price density and the price of money leads the nominal state price density, and consequently nominal bond prices, to be independent of money illusion. This article thus provides a theoretical foundation for Modigliani-Cohn’s conjecture that money illusion impacts stock markets but not bond markets.
  • Item
    The calibration of stochastic local-volatility models: an inverse problem perspective
    (Elsevier, 2019) Saporito, Yuri Fahham; Zubelli, Jorge P.
    We tackle the calibration of the Stochastic Local-Volatility (SLV) model. This is the class of financial models that combines the local volatility and stochastic volatility features and has been subject of the attention by many researchers and practitioners recently. The corresponding inverse problem consists in finding certain (functional) coefficients in a class of parabolic partial differential equations from observed values of the solutions. More precisely, given a calibrated local volatility surface and a choice of stochastic volatility parameters, we calibrate the corresponding leverage function. Our approach makes use of regularization techniques from inverse-problem theory, respecting the integrity of the data and thus avoiding data interpolation. The result is a stable and efficient algorithm which is resilient to instabilities in the regions of low probability density of the spot price and of the instantaneous variance. We substantiate our claims with numerical experiments using synthetic and real data.
  • Item
    Stochastic control with delayed information and related nonlinear master equation
    (SIAM, 2019) Saporito, Yuri Fahham; Zhang, Jianfeng (Matemático)
    In this paper we study stochastic control problems with delayed information, that is, the control at time t can depend only on the information observed before time t - h for some delay parameter h. Such delay occurs frequently in practice and can be viewed as a special case of partial observation. When the time duration T is smaller than h, the problem becomes a deterministic control problem in the stochastic setting. While seemingly simple, the problem involves certain time inconsistency issues, and the value function naturally relies on the distribution of the state process and thus is a solution to a nonlinear master equation. Consequently, the optimal state process solves a McKean--Vlasov SDE. In the general case that T is larger than h, the master equation becomes path-dependent and the corresponding McKean--Vlasov SDE involves the conditional distribution of the state process. We shall build these connections rigorously and obtain the existence of a classical solution of these nonlinear (path-dependent) master equations in some special cases.
  • Item
    Recuperação e classificação de sentimentos de usuários do twitter em período eleitoral
    (2020) Matos, Fernanda Fernandes; Magalhães, Lúcia Helena de; Souza, Renato Rocha
    Introdução: A redes sociais tornaram-se um espaço importante para usuários expressarem seus sentimentos. Esses comentários são valiosos para governantes saberem o ponto de vista dos cidadãos sobre as suas propostas políticas e candidatos perceberem a reação dos eleitores a respeito da campanha eleitoral. Objetivos: Analisar os sentimentos expressos pelos usuários no Twitter, referentes aos candidatos que concorreram à presidência do Brasil no ano de 2018, e predizer o resultado das eleições com base nessas postagens. Metodologia: Os posts sobre os candidatos que disputaram o segundo turno das eleições foram o objeto de estudo. Usou-se o software Orange Canvas, uma ferramenta de aprendizado de máquina livre e de código aberto, para a coleta da amostra e para a extração das informações relevantes. A técnica de análise de opinião foi aplicada para classificação automática dos sentimentos em positivos, negativos e neutros. Para melhor análise e interpretação dos resultados, exibiram-se as palavras mais importantes dos comentários em nuvens de palavras e as emoções, em gráficos de distribuição de frequência. Resultados: Detectaram-se muitos sentimentos negativos nas postagens e a emoção de surpresa foi a que mais se destacou para ambos os concorrentes. Conclusões: O estudo mostrou que o Twitter é um local interessante para usuários expressarem seus sentimentos no período eleitoral. Porém, o trabalho não foi capaz de prever o resultado das eleições com base nas emoções. Acredita-se que isso se deve às altas taxas de rejeição dos eleitores quanto aos candidatos e a polarização que tem caracterizado a política brasileira nos últimos tempos.
  • Item
    Representação de recursos multimídia na web: uso e reúso de padrões de anotação
    (2020) Silva, Daniela Lucas da;; Souza, Renato Rocha
    O artigo retrata os desafios impostos ao cenário de produção, organização e disseminação de informação em diversos setores da sociedade em decorrência à evolução das tecnologias para tratamento de dados multimídia na Web. Pesquisas no campo das Ciências daInformação são progressivas na busca de inovações para o universo de dados distribuídos na Web a fim de ampliar os pontos de acesso e melhorar a gestão, a organização e a recuperação de objetos digitais na rede. Nessa perspectiva, o artigo objetiva contribuir com um estudo sistemático e analítico sobre iniciativas de padrões de metadados, modelos e ontologias voltados ao domínio da descrição multimídia. Metodologicamente, a pesquisa foi classificada como sendo de natureza qualitativa e quantitativa, de caráter exploratório e descritivo à luz de literatura científica já publicada e material empírico específico, o que a torna bibliográfica e documental. O estudo culminou na obtenção de um ranking de ontologias a partir de uma análise comparativa e uma avaliação criteriosa sobre dimensões concernentes a reúso de recursos de conhecimento disponíveis na Web. Os resultados contribuem na perspectiva de possíveis soluções para o tratamento semântico de variados tipos de metadados existentes para descrição de acervos em rede que lidam com conteúdo multimídia.
  • Item
    Representação de documentos multimídia: dos metadados às anotações semânticas
    (2014) Silva, Daniela Lucas da; Rocha, Renato Souza
    O artigo apresenta o resultado do estudo do estado da arte de iniciativas internacionais e nacionais que estão explorando o uso de ontologias e padrões de metadados para agregar conhecimento em anotações a fim de realizar a organização e a integração de informações multimídia em diferentes domínios. Para tal, foi utilizado como método de pesquisa o levantamento bibliográfico e documental nos campos das ciências da informação e da computação, especialmente na área de representação documental do tipo multimídia. Permitiu-se evidenciar que há esforços e desafios no desenvolvimento de padrões de metadados, vocabulários controlados e ontologias formais na tentativa de melhor representar informações visando à recuperação semântica
  • Item
    Extraction of keywords from texts: an exploratory study using noun phrases
    (2014) Souza, Renato Rocha; Raghavan, K. S.
    The increasing use of Web for both scholarly publishing and information retrieval emphasizes the need for mechanisms to support efficient indexing and effective information retrieval. Manual indexing and knowledge representation techniques are not suitable for handling huge volumes of digital information. This paper presents an approach to extracting key phrases from texts based on the intrinsic semantics of the text. The methodology has been tested with a series of small-scale experiments involving texts in Portuguese language (SOUZA, 2005; SOUZA and RAGHAVAN, 2006). The results suggest that the approach yields satisfactory results. Some suggestions for future work have been made.
  • Item
    Specifying ubiquitous systems through the algebra of contextualized ontologies
    (2014) Cafezeiro, Isabel; Viterbo, José; Rademaker, Alexandre; Haeusler, Edward Hermann; Endler, Markus
    In this paper we present the algebra of contextualized ontologies and an approach to specify context-aware systems. The algebra is designed to support context moddeling and aims at the specification of modular and scalable description of arbitrarily complex systems. It takes contextualization as a basic notion and proposes a small set of simple and powerful operations to compose and decompose contextualized entities. The specification approach considers the gap between the formal specification and the real application and split the specification process in three levels variyng from the system design to the complete formalization using the algebra.
  • Item
    Infimum gaps for limit solutions
    (2015) Soledad Aronna, Maria; Motta, Monica; Rampazzo, Franco
    After recalling the notion of L1 limit solution for a dynamics which is affine in the (unbounded) derivative of the control, we focus on the possible occurrence of the Lavrentiev phenomenon for a related optimal control problem. By this we mean the possibility that the cost functional evaluated along L1 inputs (and the corresponding limit solutions) assumes values strictly smaller than the infimum over AC inputs. In fact, it turns out that no Lavrentiev phenomenon may take place in the unconstrained case, while the presence of an end-point constraint may give rise to an actual gap. We prove that a suitable transversality condition, here called Quick 1-Controllability, is sufficient for this gap to be avoided. Meanwhile, we also investigate the issue of trajectories’ approximation through implementation of inputs with bounded variation.
  • Item
    Usabilidade da Biblioteca Virtual em Saúde: avaliando a eficácia, eficiência e satisfação
    (2015) Lima, Izabel França de; Souza, Renato Rocha; Dias, Guilherme Ataíde
    Aborda aspectos relativos à avaliação de bibliotecas digitais, consideradas como dispositivos informacionais que podem auxiliar na democratização da informação mediada pelas tecnologias digitais. Tais bibliotecas podem ser compreendidas como um espaço de organização, armazenamento, disseminação e acesso à informação por meio de uma rede de comunicação. Discute a importância da avaliação dessas bibliotecas, observando a ausência de normas internacionais destinadas à mensuração dessas. Objetiva avaliar a usabilidade da Biblioteca Virtual em Saúde (BVS). Metodologicamente caracteriza-se como um teste formal de usabilidade, com a finalidade de medir a eficiência, a eficácia e a satisfação de usuários de bibliotecas digitais. O teste foi composto por três instrumentos de coleta de dados, um questionário de perfil/experiência; uma lista de dez tarefas a serem realizadas, utilizando o site da BVS; e um questionário com oito perguntas abertas que extraía percepções sobre o uso da biblioteca e seus recursos. O modelo metodológico foi aplicado entre os dias 05 e 21 de dezembro de 2011 no laboratório de informática do Centro de Ciências da Saúde da Universidade Federal da Paraíba. Os resultados do teste de usabilidade possibilitam inferir que a BVS apresenta um bom nível de eficácia e boa eficiência, tendo o quesito satisfação atingido o nível satisfatório, conforme as respostas apresentadas nas questões abertas pelos participantes da pesquisa. Foram detectados alguns problemas de usabilidade e apresentadas sugestões para melhorar a interface e, consequentemente, a interação usuário/biblioteca digital.
  • Item
    As wordnets do português
    (2015-04) Oliveira, Hugo Gonçalo; Paiva, Valeria de; Freitas, Cláudia; Rademaker, Alexandre; Real, Livy; Simões, Alberto
    Not many years ago it was usual to comment on the lack of an open lexicalsemantic knowledge base, following the lines of Princeton WordNet, but for Portuguese. Today, the landscape has changed significantly, and researchers that need access to this specific kind of resource have not one, but several alternatives to choose from. The present article describes the wordnet-like resources currently available for Portuguese. It provides some context on their origin, creation approach, size and license for utilization. Apart from being an obvious starting point for those looking for a computational resource with information on the meaning of Portuguese words, this article describes the resources available, compares them and lists some plans for future work, sketching ideas for potential collaboration between the projects described.