Orientador: Gabriela Stangenhaus / Dissertação (mestrado) - Universidade Estadual de Campinas, Instituto de Matematica, Estatistica e Ciencia da Computação / Made available in DSpace on 2018-07-20T01:15:10Z (GMT). No. of bitstreams: 1 Silva_DamiaoNobregada_M.pdf: 2683132 bytes, checksum: 205a0f07bbedeff4e84f86e420cb584d (MD5) Previous issue date: 1995 / Resumo: Neste trabalho será apresentada uma descrição básica de tópicos importantes do método Bootstrap (Efron, 1979a) para serem aplicados em problemas de Inferência Estatística cujas soluções analíticas são complicadas ou desconhecidas. Este texto é dirigido não só para estudantes de pós-graduação em Estatística, mas também para alunos de Bacharelado em Estatística que tenham cursado disciplinas básicas de Probabilidade e Inferência. É vista, inicialmente, a implementação do método para estimar variâncias, tendenciosidades e construção de intervalos de confiança. Em seguida, esta metodologia é implementada em problemas de análise de regressão múltipla utilizando critérios de estimação como: mínimos quadrados, norma LI e outros métodos robustos. Finalmente, é apresentada uma aplicação com dados de um experimento químico em bases de Schiff, que é um problema de regressão linear com restrição nos parâmetros, para mostrar a versatilidade do método em problemas mais complicados. / Abstract: Not informed. / Mestrado / Mestre em Estatística

Models governing stochasticity in various systems are typically calibrated from data, therefore are subject to statistical errors/uncertainties which can lead to inferior decision making. This thesis develops statistically and computationally efficient data-driven methods for problems in stochastic simulation and optimization to quantify and hedge impacts of these uncertainties. The first half of the thesis focuses on efficient methods for tackling input uncertainty which refers to the simulation output variability arising from the statistical noise in specifying the input models. Due to the convolution of the simulation noise and the input noise, existing bootstrap approaches consist of a two-layer sampling and typically require substantial simulation effort. Chapter 2 investigates a subsampling framework to reduce the required effort, by leveraging the form of the variance and its estimation error in terms of the data size and the sampling requirement in each layer. We show how the total required effort is reduced, and explicitly identify the procedural specifications in our framework that guarantee relative consistency in the estimation, and the corresponding optimal simulation budget allocations. In Chapter 3 we study an optimization-based approach to construct confidence intervals for simulation outputs under input uncertainty. This approach computes confidence bounds from simulation runs driven by probability weights defined on the data, which are obtained from solving optimization problems under suitably posited averaged divergence constraints. We illustrate how this approach offers benefits in computational efficiency and finite-sample performance compared to the bootstrap and the delta method. While resembling distributionally robust optimization, we explain the procedural design and develop tight statistical guarantees via a generalization of the empirical likelihood method. The second half develops uncertainty quantification techniques for certifying solution feasibility and optimality in data-driven optimization. Regarding optimality, Chapter 4 proposes a statistical method to estimate the optimality gap of a given solution for stochastic optimization as an assessment of the solution quality. Our approach is based on bootstrap aggregating, or bagging, resampled sample average approximation (SAA). We show how this approach leads to valid statistical confidence bounds for non-smooth optimization. We also demonstrate its statistical efficiency and stability that are especially desirable in limited-data situations. We present our theory that views SAA as a kernel in an infinite-order symmetric statistic. Regarding feasibility, Chapter 5 considers data-driven optimization under uncertain constraints, where solution feasibility is often ensured through a "safe" reformulation of the constraints, such that an obtained solution is guaranteed feasible for the oracle formulation with high confidence. Such approaches generally involve an implicit estimation of the whole feasible set that can scale rapidly with the problem dimension, in turn leading to over-conservative solutions. We investigate validation-based strategies to avoid set estimation by exploiting the intrinsic low dimensionality of the set of all possible solutions output from a given reformulation. We demonstrate how our obtained solutions satisfy statistical feasibility guarantees with light dimension dependence, and how they are asymptotically optimal and thus regarded as the least conservative with respect to the considered reformulation classes.

No description available.

At present a big part of communication passes through computer network. Amount of the communication increases every year. That is why the claims on computing power raise. The procedure of the communication processing for forensic purposes is worth paralleling to increase the computing power. Research group NES@FIT created an instrument Netfox Detective in one of its projects. From this instrument it is planned to create a distributed system for the processing of intercepted communication for the purpose of forensic analysis. Interface is one of the parts, with its help the distributed system will be operated. In my theses I will concern with the creation of the web interface for the instrument Netfox Detective which is a desktop application presently. Web interface, after little modifications, will be used for paralleled version of application. Web interface will mediate the same informations as the desktop version. To obtain information for forensic analysis it will use framework Netfox Framework identically as the desktop version. Advantage of web interface compared to the desktop version is that a user who approaches web interface will need a device with web browser. It means that a user can work with any operation system.

At present a big part of communication passes off through computer network. Amount of the communication magnifys every year. It has a consequence that claims on computing power raise. The procedure of processsing of the communication for forensic purposes pays off to parallel for the purpose of increase of computing power. Research group NES@FIT created an instrument Netfox Detective in one of its projects. From this instument it is planed to create a distibuted system for the processing of intercepted communication for the purpose of forensic analysis. Interface is one of parts with the help of its the distributed system will be operated. In my theses I will concern with the creation of the web interface for the instument Netfox Detective which is a desktop application presently. Web interface use, after little modifications, for paralleled version of application. Web interface will mediate the same informations as the desktop version. For the obtaining of informations for forensic analysis it will use framework Netfox Framework identically as the desktop version. Advantage of web interface compared to the desktop version is that user who approach to web interface need device with web browser. It means that user can work whichever operation system.

A Wald type test with the wrong dispersion matrix is used when the dispersion matrix is not a consistent estimator of the asymptotic covariance matrixof the test statistic. One class of such tests occurs when there are k groups and it is assumed that the population covariance matrices from the k groups are equal, but the common covariance matrix assumption does not hold. The pooled t test, one way AVOVA F test, and one way MANOVA F test are examples of this class. Two bootstrap confidence regions are modified to obtain large sample Wald type tests with the wrong dispersion matrix.

No description available.

Software-De_ned Networking (SDN) proposes a new network architecture inwhich the control plane and forwarding plane are decoupled. SDN can improvenetwork e_ciency and ease of management through the centralization of the controland policy decisions. However, SDN deployments are currently limited todata-center and experimental environments. This thesis surveys the deploymentof SDN from the perspective of a telecommunication network operator. We discussthe strategies which enable the operator to migrate to a network in whichboth SDN and legacy devices interoperate. As a synthesis of existing technologiesand protocols, we formulate an automated process for the bootstrapping of newlydeployed forwarding devices. Furthermore, we review solutions for programmingthe forwarding devices and for performing topology discovery. The functionalcorrectness of the proposed bootstrapping process is evaluated in an emulatedenvironment.

No description available.

The similarities between generations of living subjects are often quantified by heritability. By distinguishing genotypic variation, or variation due to parental pairings, from phenotypic variation, or normal intraspecies variation, the heritability of traits can be estimated. Due to the multivariate nature of many traits, such as size and shape, computation of heritability can be difficult. Also, assessment of the variation of the heritability estimate is extremely difficult. This study uses nonparametric methods, namely the randomization test and the bootstrap, to obtain both a measure of the extremity of the observed heritability and an assessment of the uncertainty.