The problem in this thesis was originally motivated by problems presented with documents of Early England Data Set (DEEDS). The central problem with these medieval documents is the lack of methods to assign accurate dates to those documents which bear no date. With the problems of the DEEDS documents in mind, we present two methods to impute missing features of texts. In the first method, we suggest a new class of metrics for measuring distances between texts. We then show how to combine the distances between the texts using statistical smoothing. This method can be adapted to settings where the features of the texts are ordered or unordered categoricals (as in the case of, for example, authorship assignment problems). In the second method, we estimate the probability of occurrences of words in texts using nonparametric regression techniques of local polynomial fitting with kernel weight to generalized linear models. We combine the estimated probability of occurrences of words of a text to estimate the probability of occurrence of a text as a function of its feature -- the feature in this case being the date in which the text is written. The application and results of our methods to the DEEDS documents are presented.

Making a computer &lsquo / see&rsquo / is certainly one of the greatest challanges for today. Apart from possible applications, the solution may also shed light or at least give some idea on how, actually, the biological vision works. Many problems faced en route to successful algorithms require finding corresponding tokens in different views, which is termed the correspondence problem. For instance, given two images of the same scene from different views, if the camera positions and their internal parameters are known, it is possible to obtain the 3-Dimensional coordinates of a point in space, relative to the cameras, if the same point may be located in both images. Interestingly, the camera positions and internal parameters may be extracted solely from the images if a sufficient number of corresponding tokens can be found. In this sense, two subproblems, as the choice of the tokens and how to match these tokens, are examined. Due to the arbitrariness of the image pairs, invariant schemes for extracting and matching interest points, which were taken as the tokens to be matched, are utilised. In order to appreciate the ideas of the mentioned schemes, topics as scale-space, rotational and affine invariants are introduced. The geometry of the problem is briefly reviewed and the epipolar constraint is imposed using statistical outlier rejection methods. Despite the satisfactory matching performance of simple correlation-based matching schemes on small-baseline pairs, the simulation results show the improvements when the mentioned invariants are used on the cases for which they are strictly necessary.

This thesis aims to investigate the social structure of Neolithic Period. To do this, both physical anthropological and archaeological data are used, and it is sought after whether burial customs and skeleton biology can be a parameter to understand social organization of a concerned area in a given time period. For this thesis the data comes from Abu Hureyra and &Ccedil / ay&ouml / n&uuml / . Quantified data of burial types and grave goods are used in order to create descriptive statistical graphics. Then, correspondence analysis is employed to detect statistical significance in data sets, if exists. Anthropological data is stemmed from previous researchers. On the other hand, they were still used to investigate sex and age distributions with the same tools employed before. As a final study two settlements are compared within and with each other to chase the clues for social differentiation.

This thesis examines a sample of Pitted Ware pottery from the Gotlandic site of Ajvide in regard to the spatial and temporal relationship of the shards. Ajvide Stone Age site have been extensively excavated over the past 25 years and the most recent analysis of pottery was carried out by Inger Österholm in 1987 and new analyses of the material are therefore greatly needed. Pottery from three sample areas was analysed in regard to their distribution, decoration, fragmentation, part of pot and quality of the goods. To analyse the material in a contextual way a multivariate statistical analysis called correspondence analysis was used. Two areas denominated “dark areas”, which Österholm defined previously, as possible ritual sites, and one reference area with mixed material were chosen for the analysis. By including two of these “dark areas” the study also examine if similarities or differences could be seen among the shard distribution how these areas have been defined and possibly re-defined.

Habilitationsschrift--Mannheim. / Includes indexes. Bibliography: p. [198]-205.

Made available in DSpace on 2016-06-02T19:32:16Z (GMT). No. of bitstreams: 1 DissPKA.pdf: 490127 bytes, checksum: 18f6cf8d6514765ac9c01c7df71da800 (MD5) Previous issue date: 2005-09-23 / Universidade Federal de Minas Gerais / The Cerrado Domain occupied formerly 2 million km2 of the Brazilian territory, especially in the Central Plateau. The cerrado vegetation is not uniform in physiognomy, ranging from grassland to tall woodland, but with most of its physiognomies within the range defined as tropical savanna. In cerrado, there are few areas that become waterlogged during the rainy season due to the poor drainage of the soil, allowing the appearance of a hyperseasonal cerrado, characterized by two contrasting stresses, one induced by drought and fire during the winter, the other by soil saturation in the summer. As long as soil is important in the ecology of the cerrado, limiting the cerrado distribution and the occurrence of its physiognomies, we investigated the soil-vegetation relationships in a cerrado core area in Emas National Park, in three vegetation forms: hyperseasonal cerrado, seasonal cerrado, and wet grassland. We collected vegetation and soil samples in these three vegetation forms and submitted obtained data to a canonical correspondence analysis. Our results showed a distinction among hyperseasonal cerrado, seasonal cerrado and wet grassland, which presented different floristic compositions and species abundances. The edaphic variables best related to the hyperseasonal and seasonal cerrados were sand, base saturation, pH, and magnesium. The wet grassland was related to higher concentrations of clay, organic matter, aluminium saturation, aluminium, phosphorus, and potassium. We also investigated the relationships between number of species and soil characteristics, with simple multiple linear regressions, and found that aluminium and pH were the best predictors of species density, the former positively related to species density and the latter negatively related. / O Domínio do Cerrado ocupava aproximadamente dois milhões de km2 do território brasileiro, especialmente no Planalto Central. A vegetação de cerrado não é uniforme em sua fisionomia, pois varia desde o campo limpo até o cerradão, mas a maior parte de suas fisionomias podem ser definidas como savana. Em algumas raras áreas de cerrado, a drenagem do solo é muito baixa, provocando o seu alagamento na estação chuvosa e favorecendo o surgimento de cerrados hiperestacionais, em que há dois períodos de estresse, um induzido pela seca e pelo fogo no inverno, e outro induzido pelo alagamento no verão. Como o solo é um dos principais fatores que controlam a distribuição do cerrado, além de ser determinante para a ocorrência das diferentes fisionomias que o compõem, nosso objetivo neste trabalho foi estudar a relação solo-vegetação em uma área nuclear de cerrado, no Parque Nacional das Emas (GO), em três ambientes: cerrado hiperestacional, cerrado estacional e campo úmido. Amostramos o solo e a vegetação desses três ambientes e comparamos algumas variáveis edáficas e vegetacionais por meio de análise de correspondência canônica. Nossos resultados apresentaram uma distinção florística entre o cerrado hiperestacional, o cerrado estacional e o campo úmido. As variáveis edáficas mais relacionadas com os cerrados hiperestacional e estacional foram porcentagem de areia, saturação por bases, pH e magnésio, enquanto que com o campo úmido foram argila, matéria orgânica, alumínio, saturação por alumínio e potássio. Também relacionamos o número de espécies em cada parcela com algumas variáveis edáficas, por meio de análises de regressão, e encontramos que o pH foi relacionado negativamente, e o alumínio, positivamente, com o número de espécies.

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / The present paper was carried out in two fragments of cilium Forest of Espinilho creek, located in Sant Ana do Livramento, RS state. The study was taken in a fragment of forest, without any cattle for two years, and another one, with cattle on it. The component arborous-bush and the natural regeneration in these two areas were evaluated as well as the relation of them with the environmental variants by means of a Canonic Correspondence Analysis (CCA). To do so, TRANSECTOS were set up, of variable length, distant 60 meters one from another, where there were parts of 20m x 20m for the inventory of the adult forest and subparts of 3.16m x 3.16m, inside the parts, for the inventory of the natural regeneration. In the inventory of the arborousbush vegetation, of the land with cattle and of the land without cattle, 42 and 43 parts were set up, respectively. In these parts all the live and dead subjects with CAP≥15.7cm were sampled. The collected data for the phyto-sociological analysis of the species were: CAP and total length. In the subparts, all the regenerative arborous-bush subjects higher than 0.10m and CAP>15.7cm were measured. In order to compose the matrix of the study data information on the environment of each part were collected: soil collection for the physical analysis and pH, determination of the soil depth, reading of the opening of the tree top, determination of the soil declivity and the stones existing in each land part. To obtain the phyto-sociological parameters, the vegetation data were analyzed using Mata Nativa 2® software. It was also analyzed the distribution of the subject density, which was organized according to the class of diameter and length of those two areas. To analyze the environmental correlations of gradients with the flower composition, CCA was used, using the computer program PC-ORD® for Windows, version 5.0. For the study of the arborous-bush vegetation, it is concluded that there was a higher density of subjects in the area with no cattle and, that the more representative families of both areas were the same ones: Myrtaceae e Anacardiaceae. In the area with cattle, the Ocotea acutifolia got a greater ecological importance while in the area without cattle it was the Ocotea acutifolia. Concerned to the environmental importance in the two areas, it is concluded that Ocotea acutifolia is influenced by higher soil depths, a higher level of clay and lower percentages of stones as well as more sand. It is concluded, from the analysis of regeneration, that the most abundant families were Myrtaceae e Anacardiaceae. The two areas presented a similar number of subjects in natural regeneration. However, it was observed problems of regeneration in both two areas, being these problems possibly related to the fact that the cattle step on the grass and eat the herbs. About the CCA it was noticed that the cattle did not obtain a significant correlation of the environment with the large amount of species while the area without cattle got significant results, indicating that L. molleoides e Blepharocalyx salicifolius are relates with bigger opening of dossal and a high percentage of stones, while Sebastiania commersoniana e Sebastiania brasiliensis are co-related with a lower declivity and a higher level of clay. / O presente estudo foi realizado em dois fragmentos de mata ciliar do Arroio Espinilho, localizados no município de Sant Ana do Livramento, RS. Um fragmento sem gado, há dois anos, e o outro com gado. Foram avaliados o componente arbóreo-arbustivo e a regeneração natural nas duas áreas assim como a relação das mesmas com as variáveis ambientais por meio da Análise de Correspondência Canônica (CCA). Para tanto, foram montados transectos, de comprimento variável, distantes 60 metros um do outro, onde foram alocadas parcelas de 20mx20m para o inventário da floresta adulta e subparcelas de 3,16mx3,16m, dentro das parcelas, para o inventário da regeneração natural. No inventário da vegetação arbóreoarbustiva, do fragmento com gado e do fragmento sem gado, foram montadas 42 parcelas e 43 parcelas, respectivamente. Nessas parcelas foram amostrados todos os indivíduos vivos e mortos com CAP≥15,7cm. Os dados coletados para as análises fitossociológicas das espécies foram: CAP e altura total. Nas subparcelas foram medidos todos os indivíduos regenerantes arbóreos-arbustivos com altura maior que 0,10m e CAP<15,7cm. Para compor a matriz de dados do estudo foram coletadas informações do ambiente de cada parcela: coleta de solo para realização da análise física e pH, determinação da profundidade do solo, leitura da abertura de copa, determinação da declividade do solo e pedregosidade existente em cada parcela. Para obtenção dos parâmetros fitossociológicos os dados de vegetação foram analisados utilizando o software Mata Nativa 2®. Foi analisada, também, a distribuição da densidade dos indivíduos, que foi organizada por classe de diâmetro e altura nas duas áreas. Para analisar as correlações ambientais dos gradientes com a composição florística foi utilizada a CCA, utilizando o programa computacional PC-ORD® for Windows versão 5.0. Do estudo da vegetação arbóreo-arbustiva conclui-se que houve maior densidade de indivíduos na área sem gado e que as famílias mais representativas das duas áreas foram às mesmas: Myrtaceae e Anacardiaceae. Na área com gado a Ocotea acutifolia obteve maior importância ecológica enquanto que na área sem gado foi a Lithraea molleoides. Com relação às correlações ambientais nas duas áreas concluiu-se que a O. acutifolia é influenciada por maiores profundidades de solo, maior teor de argila e menores declividades enquanto a L. molleoides é relacionada a maiores percentagens de pedregosidade e maior teor de areia. Da análise da regeneração conclui-se que as famílias mais abundantes foram a Myrtaceae e Anacardiaceae. As duas áreas apresentaram número semelhante de indivíduos na regeneração natural, no entanto, observou-se nas duas áreas problemas de regeneração sendo estes possivelmente, relacionados ao pisoteio e herbivoria do gado. Quanto a CCA verificou-se que a área com gado não obteve correlação significativa do ambiente com a abundância de espécies enquanto, a área sem gado obteve resultados significativos indicando, que a L. molleoides e Blepharocalyx salicifolius estão relacionadas com maiores aberturas de dossel e alta percentagem de pedregosidade enquanto a Sebastiania commersoniana e Sebastiania brasiliensis estão correlacionadas com menor declividade e maior teor de argila.

Made available in DSpace on 2015-03-26T13:32:16Z (GMT). No. of bitstreams: 1 texto completo .pdf: 1741383 bytes, checksum: 62a62b6f8598963cbe780a77c0cc72a2 (MD5) Previous issue date: 2012-06-28 / Correspondence analysis is a multivariate statistics technique that allows the symultaneous analysis of different categorical variables. It has a simple and intuitive aspect in the presentation of results through perceptual maps. Although correspondence analysis is available in several softwares, including free system R, the main problem that contributes to the lack of knowledge about the potential of this technique and its subsequent diffusion is the existence of several literatures of difficult comprehension, as they present linear algebra with a high level of abstraction and, additionally, for omitting important clarifying steps; on the other hand, the pure and simple reading of interpretation techniques simply become steps lacking in convincing meaning, which may lead to misunderstandings and insecurity in the interpretations and even takes away from the highest possible results. Thus, the aim of this study was to elaborate a clarifying text that showed the mathematical details of all the steps of the correspondence analysis in the form-applied method and with emphasis on its geometric construction. One of the applications of this technique with data from an experiment in the area of genetic improvement of the sugar cane and it allowed important and practical conclusions, such as, the number of talks is a key factor to anticipate the weight of the clump. This information is of the highest importance, since it can facilitate the operational aspects of selection of families for genetic enhancing, and the researcher won t need to wait for harvest season to select the families with the best clump weight traits, as they will be able to select the clump based on largest number of stalks. Therefore, it was concluded that this technique can be promising in the agrarian sciences. Even though the reading of this study takes time and disposal, it contributes to a better understanding of this technique, especially for a layman in algebraic developments, since it provides a complete vision of the step by step of the correspondence analysis through geometric vision and illustrative example. / A análise de correspondência é uma técnica estatística multivariada que permite realizar a análise simultânea de diferentes variáveis categóricas. Possui aspecto simples e intuitivo na apresentação de resultados por meio dos mapas perceptuais. Apesar da análise de correspondência estar disponível em diversos softwares, incluindo o sistema livre R, o principal problema que contribui para o desconhecimento do potencial desta técnica e de sua consequente difusão é a existência de diversos textos especializados de difícil compreensão, por apresentarem um alto nível de algebrismo, abstração e por omitirem passos elucidativos importantes. Por outro lado, a leitura pura e simples das técnicas de interpretação se torna desprovida de sentido convincente, que pode levar a equívocos e insegurança nas interpretações e impede até mesmo de tirar o maior proveito possível dos resultados. Assim, o objetivo deste trabalho foi elaborar um texto elucidativo que mostrasse os detalhes matemáticos de todas as etapas da análise de correspondência, de forma aplicada e com ênfase na construção geométrica. Uma aplicação desta técnica com dados de um experimento na área de melhoramento genético da cana-de-açúcar é apresentada na qual constatou-se que o número de colmos é um fator determinante para antecipar a categoria de peso da touceira. Esta informação é de fundamental importância, uma vez que pode facilitar os aspectos operacionais na seleção de famílias para o melhoramento genético, visto que o pesquisador não necessitará esperar a época da colheita para selecioná-las, pois poderá fazê-la previamente se baseando naquelas com os maiores números de colmos. Portanto concluiu-se que esta técnica pode ser promissora em ciências agrárias. Apesar da leitura do presente trabalho exigir tempo e disposição, o mesmo contribui para um melhor entendimento desta técnica, principalmente para um leitor não muito afeito aos desenvolvimentos algébricos, pois ele fornece uma visão completa do passo a passo da análise de correspondência simples através da visão geométrica e do exemplo ilustrativo.

My thesis investigates the dialogic aspects and literary qualities ascribed to letters during the long eighteenth century. In part this involves documenting the correspondence between letters and other genres, such as the novel. Being in correspondence encouraged writers such as Burney and Johnson to express the relationship between sender and recipient in interesting ways. I posit that the letter offered a sophisticated means for writers, including those in Richardson's circle, to represent speech and thought, and mimic (with varying degrees of indirection), that of others. I consider the editorial habits and typographical conventions that governed letter-writing during the period, honing in on Richardson's contributions. I link his claim that letters were written 'to the Moment' with broader tropes of 'occasional' style, and show how this manifests in letters' intricate modulations of tense and person. Chapter 1 details the conventions that prevailed in letters of the period, and their interactions with irony and innovation. I compare convention in the epistolary novels of Smollett and Richardson, and look at closure in the Johnson-Thrale correspondence. Chapter 2 demonstrates that various methods of combining one's voice with others were utilized in letters (such as those of the Burney family), including some that took advantage of the epistolary form and its reputation as 'talking on paper'. Chapter 3 shows the role of mimesis in maintaining the dialogic structure of letters, and links it to contemporary theories of sympathy and sentiment. Chapters 4 and 5 apply the findings about epistolary tradition, polyphony and sentimentalism to the letters of Sterne and Burns. In them, there is a mixture of sentiment and irony, and of individual and 'correspondent' styles. The conclusion discusses the editing of letters, both in situ and in preparation for publication. The twin ideals of spontaneity and sincerity, I conclude, have influenced the way we choose to edit letters in scholarly publications.

Esta dissertação centrou-se sobre a questão: De que modo o volume de capitais econômico, cultural e escolar dos vestibulandos da Universidade Federal do Rio Grande do Sul influencia suas escolhas de curso de graduação. Para responder tal pergunta foi realizado um estudo quantitativo, utilizando o software SPSS versão 13.0. Foi selecionada uma amostra de 21 cursos de graduação do vestibular da UFRGS 2007. O modelo estatístico é formado por 8 variáveis explicativas relacionadas aos capitais econômico, cultural e escolar dos vestibulandos da UFRGS. A hipótese principal da dissertação afirma que os candidatos com alto volume de capitais econômico, cultural e escolar tendem a escolher os cursos de maior prestígio, mais concorridos e que dão acesso a profissões melhor remuneradas; e candidatos com baixo volume de capitais tendem a se inscrever, ao contrário, nos cursos menos prestigiados, menos concorridos e que formam para as profissões consideradas menos rentáveis. A hipótese acessória aponta a existência de um outro fator que exerce influência na escolha do curso superior que é a estrutura e a política institucionais. Os resultados empíricos apontaram para a confirmação das duas hipóteses. Há uma relação estatisticamente comprovada entre o volume de capitais do indivíduo e a escolha do curso de graduação. No entanto, no momento da escolha está também presente o aspecto institucional, ou seja, os incentivos e constrangimentos gerados pela instituição, e sua política referente, neste caso, ao acesso e inclusão de indivíduos com origens sociais distintas. Assim, a dissertação mostrou, por exemplo, que os cursos de Direito e Administração, considerados de alto prestígio, apresentam um perfil de candidato diferente a depender do turno em que é oferecido. O turno noturno atrai candidatos cujo volume de capitais econômico, cultural e escolar é mais baixo quando comparado ao seu equivalente diurno. Deste modo, é possível afirmar que além da influência do volume de capitais econômico, cultural e escolar, há um aspecto referente à estrutura e política institucionais que também interfere na escolha do curso de graduação. / This thesis has a central question: How the economic, cultural and academic capital of university applicants from Federal University of Rio Grande do Sul affects their choices of higher education degrees. To answer that question we made a quantitative analysis relying on the software SPSS 13.0. The research selected a sample of 21 degrees of Federal University of Rio Grande do Sul. The statistical model has 8 independent variables to account for the applicants’ economic, cultural and academic capital. The main hypothesis is that applicants with a high level of economic, cultural and academic capital tend to choose the most disputed and prestigious degrees whereas applicants with a low level of capital are more likely to choose the less disputed and prestigious degrees. The second hypothesis is that there is an additional factor that has an impact on applicants’ choice of a higher education course, relative to the institutional dimension. The results confirm both hypotheses. There is a statistical correlation between the volume of economic, cultural and academic capital possessed by individuals and the choice of more prestigious degrees. However the institutional dimension is also present. For instance the study shows that Law and Administration degrees both attract different applicant profiles depending on whether classes are offered mainly at night or during daytimes. Night classes attract applicants with a lower volume of economic, cultural and academic capital as compared to applicants that choose to attend daytime classes. Therefore we can say that besides the influence of economic, cultural and academic capital there is an institutional dimension that also interferes with applicants’ choices.