Orientador: Roberto Carvalheiro / Coorientador: Danísio Prado Munari / Coorientador: Haroldo Henrique de Rezende Neves / Resumo: A análise de componentes principais (ACP) é uma técnica da estatística multivariada usada para avaliar as relações entre diferentes características a fim de eliminar a redundância resultante de suas correlações. No melhoramento genético animal, a ACP tem sido usada para explorar possíveis interpretações biológicas associadas aos componentes principais (CPs) que podem levar a caracterização de diferentes biotipos de animais. Os objetivos do presente estudo foram: i) avaliar as relações entre características de crescimento, escore visual e reprodutiva, por meio de ACP; ii) identificar, por meio de estudo de associação genômica ampla (GWAS), regiões genômicas que diferenciam os animais quanto aos diferentes componentes; e iii) avaliar a habilidade de predição de valores genéticos genômicos (GEBVs) obtidos para os CPs. Foram utilizados dados fenotípicos de 355.524 animais da raça Nelore provenientes da base de dados Aliança Nelore. Destes, foram genotipados 3.382 animais em painel lllumina® BovineHD (HD, ~777.000 SNPs) e 137 animais em painel GeneSeek Genomic Profiler Bovine HD (~76.000 SNPs). Os animais genotipados com o painel GGP-HD tiveram seus genótipos imputados para o painel mais denso (HD). Após o controle de qualidade, 3.519 animais com informações genotípicas de 471.880 SNPs permaneceram nas análises. A ACP foi realizada utilizando-se a matriz de (co)variância genética aditiva (AT) obtida a partir de análise multi-característica. As estimativas dos efeitos dos SNPs fora... (Resumo completo, clicar acesso eletrônico abaixo) / Abstract: Principal component analysis (PCA) is a multivariate statistical technique that allows evaluating relationships among different traits in order to eliminate the redundancy resulting from their correlations. In animal breeding, PCA has been used to explore possible biological interpretations associated with the principal components (PCs) that can lead to the characterization of distinguished animal's biotype. The objectives of the present study were: i) to evaluate relationships among growth, visual scores, and reproductive traits by performing a PCA; ii) to identify genomic regions associated with PCs by performing a genome-wide association study (GWAS) on the main PCs; and iii) to evaluate the prediction ability of genomic breeding values (GEBVs) obtained for the PCs. Phenotypic data from 355,524 Nellore animals provided by the Alliance Nellore database, were used in this investigation. A total of 3,382 Nellore animals were genotyped using the lllumina® BovineHD chip (HD, ~777,000 SNPs) and 137 animals were genotyped using the GeneSeek Genomic Profiler Bovine HD chip (~76,000 SNPs). The GGP-HD genotypes were imputed to the HD genotypes. After genomic data quality control, 471,880 SNPs from 3,519 animals were available. The PCA was applied on the additive genetic (co)variance matrix (AT) obtained using multi-trait analysis. For GWAS, SNP effects were estimated using the weighted single-step GBLUP and the BayesC methods. The genes identified within the top-10 ranking windows that explained the highest proportion of variance were used for further functional analyses. For the genomic prediction study, the GEBVs were predicted using three distinguish response variables: EBV of the original traits, EBV of the PCs, and EBV of a selection index used by some Nellore cattle commercial breeding programs. The geno... (Complete abstract click electronic access below) / Doutor

Orientador: Carlos Roberto Padovani / Banca: Adalberto José Crocci / Banca: Luiz Gonzaga Manzine / Resumo: Os experimentos agronômicos, em geral, apresentam uma quantidade razoável de variáveis observadas e uma complexa estrutura de variação entre e dentro dessas variáveis. Essa estrutura de variação acarreta uma dificuldade para a utilização dos procedimentos requeridos pelo modelo estatístico, em virtude do difícil acesso a programas computacionais para a análise dos dados multivariados. Uma alternativa para redimensionar a quantidade de variáveis consiste na técnica dos componentes principais, que consegue descrever um conjunto com um número menor de variáveis não correlacionadas entre si, ordenadas de maneira decrescente pelas magnitudes das variâncias, de tal forma que a variância total do conjunto inicial seja preservada. Em síntese, a prática da análise de componentes principais é considerada sob o objetivo da redução do espaço paramétrico. Uma das dificuldades encontrada pelos pesquisadores no uso da técnica dos componentes principais, consiste na determinação do número de componentes que deve ser utilizado na redução do espaço paramétrico. Dentre alguns métodos exploratórios discutidos foram apresentados quatro critérios para a escolha do número de componentes principais os quais retem de forma qualificada, a informação contida nas variáveis originais. Neste sentido, foi proposto no presente estudo, a elaboração de um programa computacional, desenvolvido em linguagem MAPLE V.3 e CLIPPER 5.1, de fácil manuseio e acessível a todos os pesquisadores das áreas agronômicas. Visando a operacionalização do aplicativo e a utilização dos procedimentos de análise multivariada, finalizou-se o estudo apresentando dois exemplos envolvendo situações observadas na literatura agronômica, onde no primeiro faz-se uma abordagem pela metodologia univariada e pela utilização de componentes principais por processo gráfico, e no segundo... (Resumo completo, clicar acesso eletrônico abaixo) / Abstract: The agronomical experiments, in general, introduce a reasonable quantity of observed variables and a variation complex structure between and within these variables. This variation structure carries a difficulty for the utilization of the procedures required by the statistical model, in view of the difficult access for computational programs for the analysis of the multivariate data. An option for redimensionate the quantity of variable consists in the technique of the principal components, which manages to describe a set with a smaller number of variable not correlated to each other, ordenate of decreasing way by the magnitudes of the variances, of such a form that the total variance of the initial set be preserved. In synthesis, the practice of the analysis of principal components is considered under the objective of the reduction of the parametric space. One of the difficulties found by the researchers in the use of the technique of the principal components, it consists in the determination of the number of components that should be used in the reduction of the parametric space. Among some argued exploratory methods were introduced four criteria for the choice of the number of principal components the ones retain of form qualified, the information contained in the original variables. In this sense, it was proposed at study present, the elaboration of a computational program, developed in language MAPLE V.3 and CLIPPER 5.1, of easy handling and accessible to all the researchers of the agronomical areas. Aiming at operationalization of the application and the utilization of the multivariate analysis procedures, it was concluded the study introducing two examples involving situations observed in the agronomical literature, where in the first an approach is done by the univariate methodology and by the utilization of principal components for prosecute graph, and in the second... (Complete abstract click electronic access below) / Mestre

A soja está entre as principais culturas mundiais, uma vez que é uma excelente fonte de proteínas e óleo. Além disso, a espécie é aproveitada também pela indústria de biocombustíveis. Considerando a importância das novas mudanças climáticas no agronegócio; para a soja, esta situação é agravada em virtude das condições de temperatura e latitude recomendadas para a semeadura. Dessa forma, para aumentar a produtividade da cultura mesmo frente ao aquecimento global, fazse fundamental o desenvolvimento de cultivares com alta produtividade e tolerantes às altas temperaturas. Neste contexto, o objetivo geral deste trabalho foi selecionar genótipos de soja tolerantes ao calor. Uma população composta por 80 PI\'s de soja e 15 testemunhas foi avaliada sob condições de altas temperaturas, com experimentos instalados nas cidades de Teresina-PI, Piracicaba-SP e Jaboticabal- SP, no ano agrícola 2013/2014. Para a avaliação dos genótipos, foram realizadas análises univariadas e multivariadas. A seleção dos genótipos mais tolerantes a altas temperaturas foi realizada via análise de componentes principais. Nas análises de variâncias univariadas, todos os caracteres mostraram efeitos de tratamentos significativos pelo teste F. Pela análise de componentes principais no experimento conduzido em Teresina-PI, os caracteres que mais contribuíram para a variabilidade dos genótipos avaliados foram: data que metade da parcela atingiu o estádio R5, altura da planta na maturidade, período de granação e valor agronômico. Em Piracicaba-SP, os caracteres que mais contribuíram para a variabilidade foram o período de granação, massa de 100 sementes e o número de dias para a maturidade. Para a seleção dos genótipos mais tolerantes ao calor em Jaboticabal- SP, considerou-se principalmente a altura e a produtividade. Para a análise de mapeamento associativo, a fenotipagem foi realizada em Teresina-PI e avaliados quatro caracteres: altura da planta na maturidade, valor agronômico, massa de cem sementes e produtividade. A genotipagem foi realizada utilizando o chip da empresa Affymetrix. O desequilíbrio de ligação entre pares de marcadores foi calculado pelo coeficiente de determinação r2 e a análise de associação entre marcadores e o fenótipo de interesse foi realizada utilizando a abordagem de modelo linear generalizado. Foram identificadas 16 associações significativas. / Soybean (Glycine max) is one of most important crops in the world. This crop is an source of protein and oil. Beyond that, the species is also utilized for the biofuels industry. The recent climate changes are important on agribusiness, the ones on soybean crop are worse than on other crops because of the conditions of temperature and latitude recommended for planting. Thus, to increase the productivity of the crop even in face of global warming, it is essential that soybean breeding programs promote the development of cultivars highly productive and tolerant to high temperatures. In this context, the aim of this study was to select genotypes for heat tolerance. A population composed of 80 soybean PI\'s and 15 experimental checks was evaluated under high temperature conditions. The experiments were conducted in the cities of Teresina-PI, Piracicaba-SP and Jaboticabal-SP, in the 2013/2014 season. For the evaluation of the genotypes, univariate and multivariate analysis were performed, and the selection of the most genotypes for heat tolerance was performed by principal component analysis (PCA). In the univariate analyzes of variance, all characters showed significant effects of treatments by test F. In the PCA in the experiment conducted in Teresina-PI, the variables that most contributed to the variability of genotypes were: date in which half of the parcel reached R5 stage, height of the plant at maturity, grain filling period and agronomic value. In Piracicaba-SP PCA, the variables that most contributed to the variability were: grain filling period, 100-grain weight and the number of days to maturity. For the selection of the most heat-tolerant genotypes in Jaboticabal-SP, the height and the yield were the variables that most contributed to the variability. In the the association mapping analysis, the genotypes were evaluated under conditions of high temperatures in Teresina-PI and evaluated for four traits: height of the plant at maturity, agronomic value, 100 grain weight and yield. The genotyping was carried out using the Affymetrix chip. The linkage disequilibrium between pairs of markers was calculated by the determination coefficient r2 and the association analysis between markers and the phenotype of interest was performed using the generalized linear model approach. A total of 16 significant marker-trait associations were detected for the four traits.

Orientador: José Roberto Môro / Banca: João Carlos de Oliveira / Banca: Antonio Sérgio Ferraudo / Banca: José Antonio Alberto da Silva / Banca: Jair Costa Nachtigal / Resumo: A pupunheira tem um potencial econômico e social muito grande, sendo a palmeira mais importante na América pré-colombiana, constituindo junto com o milho e a mandioca, a base da alimentação dos povos primitivos. Os principais produtos extraídos são o palmito e os frutos para o consumo humano direto, alimento animal, farinhas para consumo humano e óleo vegetal. Os objetivos do presente trabalho foram de utilizar uma lista de descritores morfológicos recomendada, para discriminar primeiramente as raças Pará e Putumayo e após sua validação estatística, verificar também a existência da raça Solimões, que até hoje tem sido negada. Foram aplicadas técnicas estatísticas univariadas e multivariadas na tentativa de discriminar as raças. Dos 42 descritores iniciais, 25 apresentaram diferenças significativas entre as raças e 15 tiveram aproximação normal. A análise discriminante mostrou que a raça Pará possuía 15% das plantas mal classificadas e Putumayo 14%, já com a seleção de desenvolmer para componentes principais, as percentagens foram 9 e 19%, respectivamente, para as duas raças. A população de Manacapuru, não formou grupo nas duas primeiras análises de agrupamento e nem com componentes principais. As três análises em conjunto, conseguiram discriminar as raças Pará, Putumayo e Solimões, sendo os descritores mais importantes nesta discriminação e classificação das raças: número de espigas por cacho, comprimento da ráquis, peso dos frutos, espessura das cascas, facilidade para descascar os frutos, peso das cascas, sabor dos frutos, espessura da polpa, distância morfológica dos frutos e peso das sementes. / Abstract: The peach palm has a economic and social potential very great being the palm most important in the América pre-Colombian, contribuiting together with the maize and the cassava in the indenous feeds. The target of the present work was: to use a morphological descriptor list recommended, to discriminate between two landraces and descriptors validation , to verify the existence of solimoes landraces. Univariated and multivariated statistical techniques were used to attemp discriminate the landraces. Form fort yone initial descriptors, twenty five had presented significant difference between the landraces and fifteen had presented normal approach. The discriminant analysis have showed that Pará landrace possessed fifteen percent of the plant badly c1assified and Putumayo about fourteen percent to it. In the analysis of principal component, the percentages were nine and nineteen percent, respectively, for the two landraces. Manacapuru population did not form c1usterin in the two first one analysis of and nor with principal components. Three joint analysis in the set had obtained to discriminate the Pará, Putumayo and Solimoes landraces and the discrimnant analysis with three landraces, c1assified Manacapuru of the Putumayo landrace inside. The most important descriptors in the discrimination between landraces were: numbers of ears per raceme, raquis length, fruit weight, thickness of fruits bark, facility to peel fruits, weight of fruit bark, fruit flavor, pulp thickness, morphological distance between fruits and seed weight. / Doutor

Apesar da Análise de Componentes Principais (PCA) ser um dos métodos mais importantes na análise da estrutura a termo de taxa de juros, há fortes indícios de não ser adequada para estimar fatores da curva de juros quando há presença de dependência temporal e erros de medida. Para corrigir esses problemas é indicado utilizar a matriz de covariância de longo prazo, extraindo a correta estrutura de covariância presente nestes processos. Neste trabalho, mostramos que realizar a previsão fora da amostra da curva de taxa de juros com o método de Análise de Componentes Principais (PCA) utilizando como base a matriz de covarância de longo prazo (LRCM) parece ser mais acurada comparada a PCA com base na matriz de covariância amostral. / Although Principal Component Analysis (PCA) is one of the most common methods to estimate the structure of interest rate volatility, there are strong indications that it is not adequate to estimate interest rate factors when there is temporal dependence and measurement errors. To correct these problems it is necessary to use the longterm covariance matrix, to extract the correct covariance structure present in these processes. In this work, we show that performing the out-of-sample forecasting of the interest rate curve with the Principal Component Analysis (PCA) method based on the long-term covariance matrix (LRCM) seems to be more accurate compared to PCA based on sample covariance matrix.

A manufatura de produtos customizados resulta em variedade de modelos, redução no tamanho de lotes e alternância frequente de tarefas executadas por trabalhadores. Neste contexto, tarefas manuais são especialmente afetadas por conta do processo de adaptação do trabalhador a novos modelos de produtos. Este processo de aprendizado pode ocorrer de maneira distinta dentro de um grupo de trabalhadores. Assim, busca-se o agrupamento dos trabalhadores com perfis similares de aprendizado, monitorando a formação de gargalos em linhas de produção constituídas por dissimilaridades de aprendizado em processos manuais. A presente dissertação apresenta abordagens para clusterização de trabalhadores baseadas nos parâmetros oriundos da modelagem de Curvas de Aprendizado. Tais parâmetros, os quais caracterizam o processo de adaptação de trabalhadores a tarefas, são transformados através da Análise de Componentes Principais e então utilizados como variáveis de clusterização. Na sequência, testam-se outras transformações nos parâmetros utilizando funções Kernel. Os trabalhadores são clusterizados através do método K-Means e Fuzzy C-Means e a qualidade dos agrupamentos formados é medida através do Silhouette Index. Por fim, sugere-se um índice de importância de variável baseado em parâmetros obtidos na Análise Componentes Principais com o objetivo de selecionar as variáveis mais relevantes para clusterização. As abordagens propostas são aplicadas em um processo da indústria calçadista, gerando resultados satisfatórios quando comparados a clusterizações realizadas sem a transformação prévia dos dados ou sem seleção das variáveis. / Manufacturing of customized products relies on a large menu choice, reduced batch sizes and frequent alternation of tasks performed by workers. In this context, manual tasks are especially affected by workers’ adaptation to new product models. This learning process takes place in different paces within a group of workers. This thesis aims at grouping workers with similar learning process tailored to avoid bottlenecks in production lines due to learning dissimilarities among workers. For that matter, we present a method for clustering workers based on parameters derived from Learning Curve (LC) modeling. Such parameters are processed through Principal Component Analysis (PCA), and the PCA scores are used as clustering variables. Next, Kernel transformations are also used to improve clustering quality. The data is clustered using K-Means and Fuzzy C-Means techniques, and the quality of resulting clusters is measured by the Silhouette Index. Finally, we suggest a variable importance index based on parameters derived from PCA to select the most relevant variables for clustering. The proposed approaches are applied in a footwear process, yielding satisfactory results when compared to clustering on original data or without variable selection.

Apesar da Análise de Componentes Principais (PCA) ser um dos métodos mais importantes na análise da estrutura a termo de taxa de juros, há fortes indícios de não ser adequada para estimar fatores da curva de juros quando há presença de dependência temporal e erros de medida. Para corrigir esses problemas é indicado utilizar a matriz de covariância de longo prazo, extraindo a correta estrutura de covariância presente nestes processos. Neste trabalho, mostramos que realizar a previsão fora da amostra da curva de taxa de juros com o método de Análise de Componentes Principais (PCA) utilizando como base a matriz de covarância de longo prazo (LRCM) parece ser mais acurada comparada a PCA com base na matriz de covariância amostral. / Although Principal Component Analysis (PCA) is one of the most common methods to estimate the structure of interest rate volatility, there are strong indications that it is not adequate to estimate interest rate factors when there is temporal dependence and measurement errors. To correct these problems it is necessary to use the longterm covariance matrix, to extract the correct covariance structure present in these processes. In this work, we show that performing the out-of-sample forecasting of the interest rate curve with the Principal Component Analysis (PCA) method based on the long-term covariance matrix (LRCM) seems to be more accurate compared to PCA based on sample covariance matrix.

Hyperspectral remote sensing image data, such as that recorded by AVIRIS with 224 spectral bands, provides rich information on ground cover types. However, it presents new problems in machine assisted interpretation, mainly in long processing times and the difficulties of class training due to the low ratio of number of training samples to the number of bands. This thesis investigates feasible and efficient feature reduction and image classification techniques which are appropriate for hyperspectral image data. The study is reported in three parts. The first concerns a deterministic approach for hyperspectral data interpretation. Multigroup and multiple threshold spectral coding procedures, and associated techniques for spectral matching and classification, are proposed and tested. By coding on subgroups of bands using one or three thresholds, spectral searching and matching becomes simple, fast and free of the need for radiometric correction. Modifications of existing statistical techniques are proposed in the second part of the investigation A block-based maximum likelihood classification technique is developed. Several subgroups are formed from the complete set of spectral bands in the data, based on the properties of global correlation among the bands. Subgroups which are poorly correlated with each other are treated independently using conventional maximum likelihood classification. Experimental results demonstrate that, when using appropriate subgroup sizes, the new method provides a compromise among classification accuracy, processing time and available training pixels. Furthermore, a segmented, and possibly multi-layer, principal components transformation is proposed as a possible feature reduction technique prior to classification, and for effective colour display. The transformation is performed efficiently on each of the highly correlated subgroups of bands independently. Selected features from each transformed subgroup can be then transformed again to achieve a satisfactory data reduction ratio and to generate the three most significant components for colour display. Classification accuracy is improved and high quality colour image display is achieved in experiments using two AVIRIS data sets.

This thesis presents the results of research into the use of factor models for stationary economic time series. Two basic scenarios are considered. The first is a situation where a large number of observations are available on a relatively small number variables, and a dynamic factor model is specified. It is shown that a dynamic factor model may be derived as a representation of a VARMA model of reduced spectral rank observed subject to measurement error. In some cases the resulting factor model corresponds to a minimal state-space representation of the VARMA plus noise model. Identification is discussed and proved for a fairly general class of dynamic factor model, and a frequency domain estimation procedure is proposed which has the advantage of generalising easily to models with rich dynamic structures. The second scenario is one where both the number of variables and the number of observations jointly diverge to infinity. The principal components estimator is considered in this case, and consistency is proved under assumptions which allow for much more error cross-correlation than the previously published theorems. Ancillary results include finite sample/variables bounds linking population principal components to population factors, and consistency results for principal components in a dual limit framework under a `gap' condition on the eigenvalues. A new factor model, named the Grouped Variable Approximate Factor Model, is introduced. This factor model allows for arbitrarily strong correlation between some of the errors, provided that the variables corresponding to the strongly correlated errors may be arranged into groups. An approximate instrumental variables estimator is proposed for the model and consistency is proved.

In this dissertation, application of multi-way analysis is extended into new areas of environmental chemistry, microbiology, electrochemistry and organometallic chemistry. Additionally new practical aspects of some of the multi-way analysis methods are discussed. Parallel Factor Analysis Two (PARAFAC2) is used to classify a wide range of weathered petroleum oils using GC-MS data. Various chemical and data analysis issues exist in the current methods of oil spill analysis are discussed and the proposed method is demonstrated to have potential to be employed in identification of source of oil spills. Two important practical aspects of PARAFAC2 are exploited to deal with chromatographic shifts and non-diagnostic peaks.GEneralized Multiplicative ANalysis Of VAriance (GEMANOVA) is applied to assess the bactericidal activity of new natural antibacterial extracts on three species of bacteria in different structure and oxidation forms and different concentrations. In this work while the applicability of traditional ANOVA is restricted due to the high interaction amongst the factors, GEMANOVA is shown to return robust and easily interpretable models which conform to the actual structure of the data. Peptide-modified electrochemical sensors are used to determine three metal cations of Cu2+, Cd2+ and Pb2+ simultaneously. Two sets of experiments are performed using a four-electrode system returning a three-way array of size (sample ?? current ?? electrode) and a single electrode resulting in a two-way data set of size (sample ?? current). The data of former is modeled by N-PLS and that latter using PLS. Despite the presence of highly overlapped voltammograms and several sources of non-linearity N-PLS returns reasonable models while PLS fails. An intramolecular hydroamination reaction is catalyzed by several organometallic catalysts to identify the most effective catalysts. The reaction of starting material in the presence of 72 different catalysts is monitored by UV-Vis at two time points, before and after heating the mixtures in an oven. PARAFAC is applied to the three-way data set of (sample ?? wavelength ?? time) to resolve the overlapped UV-Vis peaks and to identify the effective catalysts using the estimated relative concentration of product (loadings plot of the sample mode).