In many practical applications, it is well known that data collected inevitably contain one or more anomalous outliers; that is, observations that are well separated from the majority or bulk of the data, or in some fashion deviate from the general pattern of the data. The occurrence of outliers may be due to misplaced decimal points, recording errors, transmission errors, or equipment failure. These outliers can lead to erroneous parameter estimation and consequently affect the correctness and accuracy of the model inference. In order to solve these problems, three robust fuzzy neural networks (FNNs) will be proposed in this dissertation. This provides alternative learning machines when faced with general nonlinear learning problems. Our emphasis will be put particularly on the robustness of these learning machines against outliers. Though we consider only FNNs in this study, the extension of our approach to other neural networks, such as artificial neural networks and radial basis function networks, is straightforward. In the first part of the dissertation, M-estimators, where M stands for maximum likelihood, frequently used in robust regression for linear parametric regression problems will be generalized to nonparametric Maximum Likelihood Fuzzy Neural Networks (MFNNs) for nonlinear regression problems. Simple weight updating rules based on gradient descent and iteratively reweighted least squares (IRLS) will be derived. In the second part of the dissertation, least trimmed squares estimators, abbreviated as LTS-estimators, frequently used in robust (or resistant) regression for linear parametric regression problems will be generalized to nonparametric least trimmed squares fuzzy neural networks, abbreviated as LTS-FNNs, for nonlinear regression problems. Again, simple weight updating rules based on gradient descent and iteratively reweighted least squares (IRLS) algorithms will be provided. In the last part of the dissertation, by combining the easy interpretability of the parametric models and the flexibility of the nonparametric models, semiparametric fuzzy neural networks (semiparametric FNNs) and semiparametric Wilcoxon fuzzy neural networks (semiparametric WFNNs) will be proposed. The corresponding learning rules are based on the backfitting procedure which is frequently used in semiparametric regression.

Due to the inherently hierarchical nature of many natural phenomena, data collected rests in nested entities. As an example, students are nested in schools, school are nested in districts, districts are nested in counties, and counties are nested within states. Multilevel models provide a statistical framework for investigating and drawing conclusions regarding the influence of factors at differing hierarchical levels of analysis. The work in this paper serves as an introduction to multilevel models and their comparison to Ordinary Least Squares (OLS) regression. We overview three basic model structures: variable intercept model, variable slope model, and hierarchical linear model and illustrate each model with an example of student data. Then, we contrast the three multilevel models with the OLS model and present a method for producing confidence intervals for the regression coefficients. / text

Applied statisticians are often faced with the problem of dealing with high dimensional data sets when attempting to describe the variability of a single set of variables, or trying to predict the variation of one set of variables from another. In this study, two data reduction methods are described: Redundancy Analysis and Partial Least Squares. A hybrid approach developed by Bougeard et al., (2007) and called Continuum Redundancy-Partial Least Squares, is described. All three methods are extended to the frequency domain in order to allow the lower dimensional subspace used to describe the variability to change with frequency. To illustrate and compare the three methods, and their frequency dependent generalizations, an idealized coupled atmosphere-ocean model is introduced in state space form. This model provides explicit expressions for the covariance and cross spectral matrices required by the various methods; this allows the strengths and weaknesses of the methods to be identified.

Sparse approximation is a Bayesian inference program with a wide number of signal processing applications, such as Compressed Sensing recovery used in medical imaging. Previous sparse coding implementations relied on digital algorithms whose power consumption and performance scale poorly with problem size, rendering them unsuitable for portable applications, and a bottleneck in high speed applications. A novel analog architecture, implementing the Locally Competitive Algorithm (LCA), was designed and programmed onto a Field Programmable Analog Arrays (FPAAs), using floating gate transistors to set the analog parameters. A network of 6 coefficients was demonstrated to converge to similar values as a digital sparse approximation algorithm, but with better power and performance scaling. A rate encoded spiking algorithm was then developed, which was shown to converge to similar values as the LCA. A second novel architecture was designed and programmed on an FPAA implementing the spiking version of the LCA with integrate and fire neurons. A network of 18 neurons converged on similar values as a digital sparse approximation algorithm, with even better performance and power efficiency than the non-spiking network. Novel algorithms were created to increase floating gate programming speed by more than two orders of magnitude, and reduce programming error from device mismatch. A new FPAA chip was designed and tested which allowed for rapid interfacing and additional improvements in accuracy. Finally, a neuromorphic chip was designed, containing 400 integrate and fire neurons, and capable of converging on a sparse approximation solution in 10 microseconds, over 1000 times faster than the best digital solution.

The aim of this work is to evaluate the thermal quality of urban squares in Maceió city, AL, Brazil, and its relationship with the use of those areas and the thermal sensation of the users. The adopted procedures were the bibliographic review and a field research. In the bibliografic review was set the theoretical reference: the squares´ characteristics, the urban climate and its relationship with human thermal confort. To charachterize those squares it was used the concept of Robba and Macedo´s SQUARE (2002). As a field research after an inventory of the squares in Maceió, AL these squares were chosen and analyzed: Ricardo Lessa´s square, in Tabuleiro dos Martins district; Tenente Madalena´s square, in Cruz das Almas district; and Muniz Falcão´s square, in Ponta Verde´s district. Climate variables were measured and questionnaires were applyed, and behaviour maps were built for January and February, 2006. As thermal confort index it was used the parameters proposed by Fanger (1970). It has been found that thermal quality in square areas is an important factor for their use, specially because they refer to leasure and relaxing areas. This study represents an effort to raise questions and should continue as an investigation object. / Fundação de Amparo a Pesquisa do Estado de Alagoas / O presente trabalho tem como objetivo avaliar a qualidade térmica de praças na cidade de Maceió - AL e a sua relação com a utilização destes espaços e a sensação térmica dos usuários. Os procedimentos utilizados foram a pesquisa bibliográfica e documental e a pesquisa de campo. Na pesquisa bibliográfica e documental foi demarcado o referencial teórico: as características de praças, do clima urbano e a sua relação com o conforto térmico humano. Para caracterizar as praças foi usado o conceito de PRAÇA de Robba e Macedo (2002). Na pesquisa de campo, após a elaboração de um inventário das praças de Maceió-AL, foram escolhidas e analisadas as praças Ricardo Lessa, no Bairro Tabuleiro do Martins; Tenente Madalena, no Bairro Cruz das Almas; e Muniz Falcão, no Bairro Ponta Verde. Foram realizadas medições de variáveis climáticas, aplicados questionários e construídos mapas comportamentais nos meses de janeiro e fevereiro de 2006. Como Índice de Conforto Térmico foram utilizados os parâmetros estabelecidos por Fanger (1970). Ficou comprovado que a qualidade térmica dos espaços nas praças é um importante fator para a sua utilização, principalmente quando se trata de uma área destinada ao lazer e descanso. Em virtude dos limites e dificuldades, este estudo representa esforço de reflexão e um levantamento de questões que devem continuar sendo objeto de investigação.

The term mobility is at the same time, cause and effect of urban development and is directly related to the economic, social and environmental aspects of a region. Your qualified plan implies harmonic, efficient and democratically run cities. Whereas parking lots are defined as infrastructure for urban mobility, the research objective was to identify the perceptions of users of municipal services paid parking lots in order to provide subsidies for the construction of the Mobility Plan Territorial Bagé. Therefore, the present study uses the model ECSI (European Customer Satisfaction Index) to measure relationships involving users of the municipal system of rotating Paid parking. ECSI was estimated by PLS-PM method that is characterized by its robustness in the face of structural models with missing data with normality. Data were collected through questionnaires and applied, non-randomly by 401 users Municipal Parking System Revolving in the city of Bagé, Rio Grande do Sul. The results indicate that the system user considers the level of service provided satisfactory (mean 7.65), and this construct most affected by the expectation of users and image of service. The construct with the lowest average was the perceived value (average 6.73). The constructs differ systematically only for income and age profiles, in other words, as age or income increase, the trend is that there is a better evaluation of the different constructs. / O termo mobilidade é, ao mesmo tempo, causa e efeito do desenvolvimento urbano e está diretamente relacionado com os aspectos econômicos, sociais e ambientais de uma região. O seu planejamento qualificado implica cidades harmônicas, eficientes e democraticamente geridas. Considerando que estacionamentos são definidos como infraestrutura de mobilidade urbana, o objetivo da pesquisa foi o de identificar a percepção dos usuários do serviço municipal de estacionamento rotativo a fim de fornecer subsídios para a construção do Plano de Mobilidade Territorial de Bagé. Para tanto, o presente estudo utiliza o modelo ECSI (Índice Europeu de Satisfação do Cliente) na mensuração das relações que envolvem os usuários do sistema municipal de estacionamento rotativo Pago. O ECSI foi estimado pelo método PLS-PM que se caracteriza por sua robustez diante de modelos estruturais compostos por dados com falta de normalidade. Os dados foram coletados mediante questionários aplicados e respondidos, de forma não aleatória, por 401 usuários do sistema municipal de estacionamento rotativo na cidade de Bagé, Rio Grande do Sul. Os resultados obtidos permitem concluir que o usuário do sistema considera o nível de serviço prestado satisfatório (média 7,65), sendo este constructo mais afetado pela expectativa dos usuários e imagem do serviço prestado. O constructo com menor média foi o valor percebido (média 6,73). Os constructos diferem de forma sistemática apenas para os perfis renda e idade, ou seja, à medida que a idade ou a renda aumentam, a tendência é de que exista uma melhor avaliação dos diferentes constructos.

O presente trabalho propõe um procedimento de valoração de marca a partir de uma modelagem por equações estruturais. Este estudo aponta inicialmente um modelo conceitual de valoração de marca baseado na linha comportamental, ou seja, nas percepções dos consumidores e não no valor monetário da marca. Para aplicação do procedimento proposto, foram escolhidas as marcas das operadoras de telefonia celular Vivo e Claro. A partir da modelagem por equações estruturais, ajustou-se o modelo teórico proposto a cada uma das marcas em questão. Com o modelo ajustado foi possível calcular não apenas o valor de marca para cada um dos consumidores, como todas as variáveis que o compõe. A possibilidade de se calcular o valor de marca para cada consumidor é extremamente inovadora, uma vez que nenhum dos modelos revistos na literatura científica aponta esta funcionalidade. Com a base de dados contendo o valor de marca para cada consumidor entrevistado, foi possível efetuar diversas análises estatísticas, que permitiram compreender de forma profunda quais variáveis mais impactam o valor das marcas, ou ainda quais são seus pontos fortes e fracos. Com isto, foi possível desenhar estratégias de marketing específicas para que cada uma das marcas aumente o seu valor. / The present study brings a procedure that measures brand equity using a structural equation modeling. This thesis proposes a conceptual model that measures brand equity based in a behaviorist approach, i.e., based on the consumers perception instead of monetary brand value. To apply this procedure two cellular phone operator brands were chosen: Vivo and Claro, and then, using structural equation modeling, it was possible to adjust the initial theory model to each brand. From the adjusted model it was possible to measure not only brand equity to each consumer, but all variables that are part of it. The possibility to measure brand equity to each consumer is extremely new, once no other models - reviewed in the scientific literature - pointed this feature. Using the database and the brand value to each interviewed consumer it was possible to make statistic analysis that allowed comprehending deeply which variables impact brand equity, or what are their strong and weak points. Then it was possible to make specific marketing strategies to each brand to increase their equity.

This thesis is concerned with data analysis, especially with principal component analysis and its sparse modi cation (SPCA), which is NP-hard-to- solve. SPCA problem can be recast into the regression framework in which spar- sity is usually induced with ℓ1-penalty. In the thesis, we propose to use iteratively reweighted ℓ2-penalty instead of the aforementioned ℓ1-approach. We compare the resulting algorithm with several well-known approaches to SPCA using both simulation study and interesting practical example in which we analyze voting re- cords of the Parliament of the Czech Republic. We show experimentally that the proposed algorithm outperforms the other considered algorithms. We also prove convergence of both the proposed algorithm and the original regression-based approach to PCA. vi

This thesis describes methods of secret communication based on latin squares and their close relative, quasigroups. Different types of cryptosystems are described, including ciphers, public-key cryptosystems, and cryptographic hash functions. There is also a chapter devoted to different secret sharing schemes based on latin squares. The primary objective is to present previously described cryptosystems and secret sharing schemes in a more accessible manner, but this text also defines two new ciphers based on isotopic latin squares and reconstructs a lost proof related to row-latin squares. / Denna uppsats beskriver kryptosystem och metoder för hemlighetsdelning baserade på latinska kvadrater och det närliggande konceptet kvasigrupper. Olika sorters chiffer, både symmetriska och asymmetriska, behandlas. Dessutom finns ett kapitel tillägnat kryptografiska hashfunktioner och ett tillägnat metoder för hemlighetsdelning. Huvudsyftet är att beskriva redan existerande metoder för hemlig kommunikation på ett mer lättillgängligt sätt och med nya exempel, men dessutom återskapas ett, till synes, förlorat bevis relaterat till rad-latinska kvadrater samt beskrivs två nya chiffer baserade på isotopa latinska kvadrater.

No description available.