Orientador: Hildete Prisco Pinheiro / Dissertação (mestrado) - Universidade Estadual de Campinas, Instituto de Matemática, Estatística e Computação Científica / Made available in DSpace on 2018-08-17T18:30:16Z (GMT). No. of bitstreams: 1 Cuyabano_BeatrizCastroDias_M.pdf: 2595939 bytes, checksum: 953e397b509acd3e1a11de6f0e8c015b (MD5) Previous issue date: 2011 / Resumo: Modelos Multivariados são propostos para descrever a frequência de códons em sequências de DNA, bem como a ordem e frequência em que as bases nitrogenadas se apresentam em cada códon, considerando a dependência entre as bases dentro do códon. Modelos logísticos regressivos são utilizados com diferentes estruturas de dependência entre as posições do códon. Também, modelos baseados em uma extensão da representação de Bahadur para o caso multinomial são propostos para explicar dados multinomiais correlacionados. Uma aplicação desses modelos para o gene NADH4 do genoma mitocondrial humano é apresentada, e comparações desses modelos são feitas a partir de diferentes critérios como AIC, BIC e validação cruzada. Por fim, uma breve análise de diagnósticos é realizada para os modelos logísticos regressivo / Abstract: Multivariate models are proposed to describe the codons frequencies in DNA sequences, as well as the order and frequency that nucleotide bases have in each codon, considering the dependence among the bases inside a codon. Logistic regressive models are used with different structures of dependence among the three positions in a codon. Also, models based on a multinomial extension of the Bahadur's representation are proposed to explain correlated multinomial data. An application of these models to the NADH4 gene from human mitochondrial genome is presented, and model comparisons among them are done by different criteria such as AIC, BIC and cross validation. At last, a brief diagnostic analysis is done upon the logistic regressive models / Mestrado / Estatistica / Mestre em Estatística

Orientador: Aurelio Ribeiro Leite de Oliveira / Dissertação (mestrado) - Universidade Estadual de Campinas, Instituto de Matemática, Estatística e Computação Científica / Made available in DSpace on 2018-08-18T09:56:13Z (GMT). No. of bitstreams: 1 Contharteze_Eliana_M.pdf: 964639 bytes, checksum: d9a9a761fbb64f1950cd4f206057d779 (MD5) Previous issue date: 2011 / Resumo: Os métodos de pontos interiores primais-duais para o problema de regressão pela norma-p desenvolvidos terão sua estrutura matricial resultante explorada objetivando uma implementação eficiente. O problema de regressão tem inúmeras aplicações em diversas áreas. A norma-2 é muito popular, entre outros motivos, por permitir uma solução direta. Por sua vez a norma-1 permite reduzir o efeito de pontos discrepantes enquanto que a norma infinito garante proteção contra o pior caso. A norma-p permite pensar estas características de diferentes formas, adaptando o método ao problema a ser resolvido. A implementação dos métodos de pontos interiores a ser desenvolvida será comparada com implementações eficientes de outros métodos já existentes / Abstract: The specific primal-dual interior point methods for the p-norm fitting problem previously developed will have their resulting matrix structure exploited aiming at an efficient implementation. The fitting problem has numerous applications in various areas. The 2-norm is very popular, among other reasons, for allowing a direct solution. The 1-norm allows the reduction of the effect of outliers while the infinite norm provides protection against the worst case. The p-norm allows to think these characteristics in different ways adapting the method to the problem to be solved. The interior point method implementation to be developed will be compared with an efficient implementation of other existing methods / Mestrado / Matematica Aplicada / Mestre em Matemática Aplicada

Em pesquisa Genômica é de grande interesse o mapeamento de genes que controlam traços ou fenótipos quantitativos. Metodologias estatsticas para identicar genes que tenham efeitos sobre um unico traço são bem conhecidas na literatura e têm sido exaustivamente aplicadas no mapeamento genético de muitas doenças. Porem, na pratica, diferentes traços são correlacionados, como é o caso de hipertensão e obesidade, possivelmente, devido a aço de genes comuns envolvidos na sua regulação. Nestes casos, por meio de tecnicas estatísticas multivariadas, que exploram a estrutura de covariância entre os traços, é possvel identificar genes não detectados por analises univariadas, ganhar precisão nas estimativas dos efeitos e conhecer a posicão desses genes, alem de testar efeitos de pleiotropia (um mesmo gene controlando varios traços) e interacções gene-ambiente (os genes que controlam a pressão antes e depois de dieta com sal). Neste trabalho diferentes alternativas de analise estatstica são consideradas para explorar a informacão de vários tracos conjuntamente: modelo de regressão intervalar multivariado (Jiang & Zeng, 1995), mapeamento multivariado via a teoria espectral (Mangin et al.,1998), via medidas resumo relevantes (como a diferenca entre respostas antes e depois de uma exposição) e via ajustes por covariaveis. Também são introduzidas algumas abordagens graficas para o estudo do efeito de pleiotropia e interação geneambiente. As metodologias supracitadas são aplicadas a dados reais fornecidos pelo Laboratorio de Cardiologia e Genética Molecular do InCor/USP, que consideram várias medidas de pressão arterial em ratos provenientes de uma população F2. / In Genomic research, the mapping of genes which control quantitative traits has been of great interest. Statistical methods for detection of genes, in uencing a single trait, are well known in the literature and they have been exhaustive used in the genetic mapping of many diseases. However, in real situations, dierent kind of traits are correlated, such as hypertention and obesity, that would be due to the action of a set of commom genes involved in the regulation of these traits. In these cases, through of multivariate statistical techniques, which explore the covariance structure between the traits, it is possible to identify genes that are not detected by univariated analysis. In addition multivariate analysis are useful to obtain accurate estimates and to know the position of these genes, besides testing eects of pleiotropic (a gene controlling several traits) and geneenvironmental interations (genes that control the pressure before and after salt diet). In this work dierent alternatives from statistical analysis are considered to explore information of several traits jointly: Interval multivariate regression models (Jiang and Zeng, 1995); multivariate mapping through the espectral theory (Mangin et al. 1998), summary measures (for example, models formulated in terms of the dierence between two traits) and adjustments including covariates. Also, graphics procedures are introduced in order to study eects of pleiotropy and geneenvironmental interactions . The methodologies mentioned above are applied to real data set, supplied by the Cardiology and Molecular Genetic Laboratory of Heart institute (InCor-USP), that consider several measurements of blood pressure in rats that come from a F2 population.

Dados composicionais são constituídos por vetores cujas componentes representam as proporções de algum montante, isto é: vetores com entradas positivas cuja soma é igual a 1. Em diversas áreas do conhecimento, o problema de estimar as partes $y_1, y_2, \\dots, y_D$ correspondentes aos setores $SE_1, SE_2, \\dots, SE_D$, de uma certa quantidade $Q$, aparece com frequência. As porcentagens $y_1, y_2, \\dots, y_D$ de intenção de votos correspondentes aos candidatos $Ca_1, Ca_2, \\dots, Ca_D$ em eleições governamentais ou as parcelas de mercado correspondentes a industrias concorrentes formam exemplos típicos. Naturalmente, é de grande interesse analisar como variam tais proporções em função de certas mudanças contextuais, por exemplo, a localização geográfica ou o tempo. Em qualquer ambiente competitivo, informações sobre esse comportamento são de grande auxílio para a elaboração das estratégias dos concorrentes. Neste trabalho, apresentamos e discutimos algumas abordagens propostas na literatura para regressão sobre dados composicionais, assim como alguns métodos de seleção de modelos baseados em inferência bayesiana. \\\\ / Compositional data consist of vectors whose components are the proportions of some whole. The problem of estimating the portions $y_1, y_2, \\dots, y_D$ corresponding to the pieces $SE_1, SE_2, \\dots, SE_D$ of some whole $Q$ is often required in several domains of knowledge. The percentages $y_1, y_2, \\dots, y_D$ of votes corresponding to the competitors $Ca_1, Ca_2, \\dots, Ca_D$ in governmental elections or market share problems are typical examples. Of course, it is of great interest to study the behavior of such proportions according to some contextual transitions. In any competitive environmet, additional information of such behavior can be very helpful for the strategists to make proper decisions. In this work we present and discuss some approaches proposed by different authors for compositional data regression as well as some model selection methods based on bayesian inference.\\\\

Submitted by Renata Lopes (renatasil82@gmail.com) on 2016-07-15T19:25:58Z No. of bitstreams: 1 franklinsoldati.pdf: 1201476 bytes, checksum: 7d6ed9cdb8a1cf797f73ed5eee91d1a2 (MD5) / Approved for entry into archive by Adriana Oliveira (adriana.oliveira@ufjf.edu.br) on 2016-07-22T14:59:02Z (GMT) No. of bitstreams: 1 franklinsoldati.pdf: 1201476 bytes, checksum: 7d6ed9cdb8a1cf797f73ed5eee91d1a2 (MD5) / Made available in DSpace on 2016-07-22T14:59:02Z (GMT). No. of bitstreams: 1 franklinsoldati.pdf: 1201476 bytes, checksum: 7d6ed9cdb8a1cf797f73ed5eee91d1a2 (MD5) Previous issue date: 2011-02-28 / Este estudo pretendeu comprovar o maior comparecimento do eleitorado juizforano em relação ao de outras importantes cidades brasileiras nas eleições para o executivo. Ele se mostrou conclusivo para as eleições presidenciais e estaduais de 2006 e não ocorreu o mesmo para as eleições municipais de 2008. A discussão se volta para aqueles estudos que consideram a predominância, no Brasil, de um eleitorado fraco, que privilegia as elites como as reais detentoras das “rédeas” do desenvolvimento material e intelectual das sociedades e enfatiza os poderes da representação, que teriam, de uma forma ou de outra, a mídia como seu representante mais ilustre. Também analisou-se aqueles estudos que defendem um eleitorado forte, cuja participação política é ativa, mesmo se restrita à participação eleitoral, e defendem a ideia de que o eleitor quando participa de sua vida social, ouve, discute, pondera, reflete e vota. Para o confronto entre tais oposições (eleitorado fraco / eleitorado forte), foram avaliados os dados levantados das eleições citadas. A técnica da Análise Fatorial foi utilizada de modo a criar variáveis indicadoras que conseguissem sintetizar um conjunto de variáveis socioeconômicas do Atlas do Desenvolvimento Humano do Brasil do IBGE, 2000. Essas variáveis foram tomadas como independentes. Também realizou-se uma bateria de modelos de regressão, onde as variáveis eleitorais, extraídas do site do TSE, assumiram a função de variáveis dependentes. Essas análises adotam como suporte teórico uma tradição que defende a educação como porta de entrada à participação política, em geral, e eleitoral em particular, bem como reafirmar uma das mais importantes interpretações do processo eleitoral, seu significado político. O estudo culmina com a proposta do conceito de um “Perfil Urbano Brasileiro” composto pelos seus dois eixos ortogonais: o “Movimento de Reafirmação Urbana” e a “Preocupação Social Urbana” como método de avaliação crítica da infraestrutura urbana das cidades brasileiras, perfil que tende a determinar o comparecimento do eleitorado nas eleições para o executivo no Brasil. / This study aimed at proving a major attendance of the Juiz de Fora electorate in relation to others important Brazilians cities electorate in the elections to the Executive. This demonstrated itself conclusive to the Brazilian presidential and state elections in 2006, but the same did not occur to the municipal elections of 2008. The discussion turns to those studies that consider the predominance of a weak electorate, in Brazil, that benefits the elites as the real holders of the material and intellectual societies’ development control, and emphasizes the representation powers, which would have, in a way or another, the media as their most famous representative. It was also analyzed those studies that advocate a strong electorate, whose political participation is active, even it is restricted to the electoral participation, and that defend the idea that the elector when participates of his/her social life, listens, discusses, meditates, thinks and votes. In order to the confrontation of such oppositions (weak electorate/strong electorate), the data raised from mentioned elections were evaluated. The Factorial Analysis Technique was used in order to create indicator variables which were able to synthesize a set of socio-economic variables from the IBGE’s Atlas of the Brazil Human Development, 2000. These variables were taken as independents and a suite of regression models was also performed, in which the electoral variables, extracted from the TSE web site, assumed the dependence variables function. These analyses assumed as theoretical support, a tradition that defends education as the entrance door to the political participation, in general, and electoral in private; as well reaffirming one of the most important interpretations of the electoral process, its political meaning. The study culminates in the proposal of a “Brazilian Urban Profile” concept, and of its two orthogonal axes the “Urban Reaffirmation Movement” and the “Urban Social Preoccupation” as method of critical evaluation of the Brazilian cities’ urban infra-structure, which tends to determine the electorate attendance to the elections to the executive in Brazil.

abstract: Functional regression models are widely considered in practice. To precisely understand an underlying functional mechanism, a good sampling schedule for collecting informative functional data is necessary, especially when data collection is limited. However, scarce research has been conducted on the optimal sampling schedule design for the functional regression model so far. To address this design issue, efficient approaches are proposed for generating the best sampling plan in the functional regression setting. First, three optimal experimental designs are considered under a function-on-function linear model: the schedule that maximizes the relative efficiency for recovering the predictor function, the schedule that maximizes the relative efficiency for predicting the response function, and the schedule that maximizes the mixture of the relative efficiencies of both the predictor and response functions. The obtained sampling plan allows a precise recovery of the predictor function and a precise prediction of the response function. The proposed approach can also be reduced to identify the optimal sampling plan for the problem with a scalar-on-function linear regression model. In addition, the optimality criterion on predicting a scalar response using a functional predictor is derived when the quadratic relationship between these two variables is present, and proofs of important properties of the derived optimality criterion are also provided. To find such designs, an algorithm that is comparably fast, and can generate nearly optimal designs is proposed. As the optimality criterion includes quantities that must be estimated from prior knowledge (e.g., a pilot study), the effectiveness of the suggested optimal design highly depends on the quality of the estimates. However, in many situations, the estimates are unreliable; thus, a bootstrap aggregating (bagging) approach is employed for enhancing the quality of estimates and for finding sampling schedules stable to the misspecification of estimates. Through case studies, it is demonstrated that the proposed designs outperform other designs in terms of accurately predicting the response and recovering the predictor. It is also proposed that bagging-enhanced design generates a more robust sampling design under the misspecification of estimated quantities. / Dissertation/Thesis / Doctoral Dissertation Statistics 2020

abstract: Longitudinal studies contain correlated data due to the repeated measurements on the same subject. The changing values of the time-dependent covariates and their association with the outcomes presents another source of correlation. Most methods used to analyze longitudinal data average the effects of time-dependent covariates on outcomes over time and provide a single regression coefficient per time-dependent covariate. This denies researchers the opportunity to follow the changing impact of time-dependent covariates on the outcomes. This dissertation addresses such issue through the use of partitioned regression coefficients in three different papers. In the first paper, an alternative approach to the partitioned Generalized Method of Moments logistic regression model for longitudinal binary outcomes is presented. This method relies on Bayes estimators and is utilized when the partitioned Generalized Method of Moments model provides numerically unstable estimates of the regression coefficients. It is used to model obesity status in the Add Health study and cognitive impairment diagnosis in the National Alzheimer’s Coordination Center database. The second paper develops a model that allows the joint modeling of two or more binary outcomes that provide an overall measure of a subject’s trait over time. The simultaneous modelling of all outcomes provides a complete picture of the overall measure of interest. This approach accounts for the correlation among and between the outcomes across time and the changing effects of time-dependent covariates on the outcomes. The model is used to analyze four outcomes measuring overall the quality of life in the Chinese Longitudinal Healthy Longevity Study. The third paper presents an approach that allows for estimation of cross-sectional and lagged effects of the covariates on the outcome as well as the feedback of the response on future covariates. This is done in two-parts, in part-1, the effects of time-dependent covariates on the outcomes are estimated, then, in part-2, the outcome influences on future values of the covariates are measured. These model parameters are obtained through a Generalized Method of Moments procedure that uses valid moment conditions between the outcome and the covariates. Child morbidity in the Philippines and obesity status in the Add Health data are analyzed. / Dissertation/Thesis / Doctoral Dissertation Statistics 2020

No description available.

Functional data analysis (FDA) has grown into a substantial field of statistical research, with new methodology, numerous useful applications and interesting novel theoretical developments. My dissertation focuses on the empirical properties of functional regression models and their application to financial data. We start from testing the empirical properties of forecasts with the functional autoregressive models based on simulated and real data. We define intraday returns and consider their prediction from such returns on a market index. This is an extension to intraday data of the Capital Asset Pricing model. Finally we investigate multifactor functional models and assess their suitability for the prediction of intraday returns for various financial assets, including stock and commodity futures.

No description available.