O estudo da variação da composição atmosférica contribui para a compreensão dos efeitos das emissões antrópicas sobre a troposfera. A análise química da água de chuva assume importância neste contexto, uma vez que a precipitação úmida constitui um excelente meio de remoção de muitos poluentes da atmosfera. O aumento de acidez das chuvas, em várias partes do planeta, tem sido associado, predominantemente, à presença de ácidos fortes como o sulfúrico e nítrico, e ácidos orgânicos como acético e fórmico, originados pela oxidação de compostos de enxofre, nitrogênio e carbono, respectivamente. Dando continuidade a pesquisas anteriores sobre deposição úmida realizadas no IQ-USP, no presente trabalho buscou-se extrair mais informações de titulações potenciométricas e condutométricas de água de chuva, mediante aperfeiçoamento dos métodos de cálculo, com recursos quimiométricos modernos, de modo a complementar as medidas de pH e condutividade convencionais. As concentrações de espécies ácidas ou básicas nas amostras de água de chuva são da ordem de 10-5 mol L-1, muito abaixo das concentrações empregadas na prática cotidiana das titulações potenciométricas (tipicamente, no intervalo de 10-1 a 10-3 mol L-1). Dessa forma, propôs-se o uso de um método de regressão não linear (RNL) multiparamétrica, baseado no método de Levenberg-Marquardt, para a análise de dados potenciométricos (TP-RNL) após procedimento de transposição de coordenadas das curvas de titulação. Ao lado das determinações das concentrações dos analitos, pôde-se estimar ou refinar as constantes de protonação das bases de Brönsted presentes. Para tanto, procedeu-se ao ajuste por RNL das concentrações e/ou das constantes de protonação das bases presentes, de modo a minimizar o quadrado dos desvios entre os valores experimentais e os calculados em função do titulante adicionado, recorrendo a uma equação geral baseada no número médio de protonação de cada uma das bases. Ficou demonstrada a possibilidade de se determinar a soma de ácidos fortes completamente dissociados, soma de ácidos fracos com pK próximo a 4 (principalmente ácidos acético e fórmico), HCO3-/H2CO3 e a soma de bases fracas com pK ao redor de 9 (essencialmente NH3/NH4+). Durante as titulações, a facilidade em se coletar medidas de condutância simultaneamente às de pH levou a elaboração de um método sofisticado de análise de curvas de titulação condutométrica. Para tanto, desenvolveu-se uma equação de RNL baseada na contribuição das espécies em cada valor de pH e na condutância iônica medida (TC-RNL). A equação de TC-RNL é alimentada por dados de pH em função do titulante adicionado, oriundos de titulações potenciométricas. Dessa forma, foi possível caracterizar diretamente amônio e bicarbonato, além do conteúdo iônico total que não participa dos equilíbrios ácidos/base existentes no sistema. Adicionalmente, foi possível separar a contribuição do ácido acético da do fórmico, tarefa impossível por TP-RNL, ao menos nas baixas concentrações encontradas na chuva. Nas mesmas amostras de chuva tituladas foi procedida a análise por eletroforese capilar de zona com detecção condutométrica sem contato (CZE-CCD), conforme procedimento descrito na literatura, estimando-se os cátions e ânions majoritários. Num aperfeiçoamento do método para ânions, viabilizou-se a determinação simultânea HCO3-. Para validar os resultados obtidos por CZE-CCD, recorreu-se à cromatografia de íons. As concentrações de vários analitos, dependendo das amostras de chuva, avizinham-se do limite de quantificação das técnicas TP-RNL, TC-RNL e CZE-CCD. Mesmo assim, para as amostras de chuva analisadas, em geral, houve concordância satisfatória entre os valores estimados pelas três técnicas para espécies agrupadas da seguinte forma: ácidos fortes, ácidos fracos, amônio e bicarbonato. / Ivestigation of the variability of atmospheric chemical composition is an important tool for the evaluation and comprehension of anthropic emissions effects.Rainwater analysis has great importance in this context because wet deposition represents an efficient removal route of many atmospheric pollutants. The acidification of rainwater over wide regions of the planet is associated with the increased presence of strong acids, like sulfuric and nitric ones, and organic acids, such as acetic and formic, mainly originated from oxidative reactions od sulfur, nitrogen and carbon compounds in the atmosphere. In the context of a broad research line at IQ-USP devoted to the study of wet deposition and associated atmospheric chemistry, this dissertation advances the subject of using moderm chemometric methods to extract reliable analytical information from potentiometric and conductometric titrations of rainwater. The concentrations of acidic and alkaline species in rainwater are typically below 5 x 10-5 mol L-1, out of the reach of conventional potentiometric titrations (widely applied in the range of 10-1 and 10-3 mol L-1). Thus, multiparametric non-linear regression based on Levenberg-Marquardt method combined with coordinates transposition to reduce iterative calculation is proposed for the analysis of potentiometric data (PT-NLR). General equations based on the average protonation number of each base are used to describe the system. Bases concentrations and/or protonation constants are adjusted by NLR in order to minimize the sum-of-squares of the residuals between experimental and calculated values of the titration curves. The data analysis was tuned to provide four parameters as output: the sum of strong acids, the sum of weak acids with pK around 4.2 (mainly acetic and formic acids), the sum of HCO3-/H2co3 and the sum of weak bases with pK near 9 (essentially NH3/NH4+). To extract more analytical information from the acid/base titrations, conductance meassurements were made simultaneously with the pH readings, and a sophisticated data analysis method was envisioned. The pH data and available protonation constants for bases included in the model are used to compute the distribution of the species. The available conductivity of each species of the model is introduced in a general equation where the only unknowns are the total concentrations of the bases and of electrolyte not involved in acid/base equilibria. All these concentrations are adjusted by NLR on the conductance data, resulting in the first successful conductometric titration method with non-linear regression analysis, CT-NLR. The great advantage over PT-NLR was the capability to distinguish the contributions of acetic and formic acids in rainwater (an impossible task in PT-NLR method), but total strong acids, ammonium and bicarbonate are determined as well, beside the total ionic content of species not involved acid-base equilibria. Rainwater samples were also analysed by capillary zone electriphoresis equipment with contactless conductometric detection (CZE-CCD), as decribed in literature, regarding the major cations and anions. An improved CE method for anions was developed to afford the simultaneous determination of HCO3-. Validation of results was made with the reference method of ion chromatography. Although in many samples some analytes are low and close to the quantification limits od CZE-CCD, PT-NLR and CT-NLR, satisfactory agreement was achieved between values obtained by the three techniques for strong acids, weak acids, ammonium and bicarbonate.

A uniformidade de emissão ao longo de uma linha lateral de irrigação localizada em nível, está relacionada com a variação de pressão, ocasionada pela perda de carga durante o escoamento da água no tubo emissor e pela saída de água pelos emissores ao longo da linha lateral. Para calcular o comprimento máximo das linhas laterais, é necessário determinar de forma precisa essa variação de pressão para atender um critério pré-estabelecido de variação de vazão. O objetivo desse trabalho foi desenvolver metodologia para o cálculo da perda de carga em linhas laterais de gotejamento com emissores do tipo bóbi, para tal foram desenvolvidos dois modelos, um usando regressão linear múltipla e outro usando análise dimensional. Ambos tiveram como base equações de perdas de carga contínua no tubo e localizadas nos emissores disponíveis na literatura. As variáveis dessas equações ficaram definidas como variáveis de entrada para os modelos e foram determinadas em laboratório, usando 12 modelos de tubos emissores, com diferentes diâmetros, espaçamentos entre emissores e seções transversais médias dos emissores. Os dois modelos apresentaram ótimo ajuste aos dados observados, com coeficiente de determinação ajustado de 0,977 e 0,99 para os modelos obtidos a partir de regressão linear múltipla e a partir de análise dimensional, respectivamente. O modelo obtido a partir de análise dimensional considera a variação de diâmetro que o tubo sofre com a pressão da água. Ambos os modelos são de fácil uso. / The uniformity of emission along a lateral line of drip irrigation, is related of the pressure variation caused by head losses during the flow in dripline and the water outlet by emitters along the lateral line. To calculate the maximum length of lateral lines, we need to accurately determine the variation of pressure to meet a pre-established criterion of flow variation. The objective of this work was to develop methodology for calculating the head loss in driplines with emitters type \"bóbi\". For this we developed two models, one using multiple linear regression and one using dimensional analysis. Both were based on equations of continuos head losses in the tube and located in the emitters available in the literature. The equations of these variables were defined as input variables for the models and were determined in the laboratory, using 12 models driplines with different diameters, spacing between emitters and emitters of medium cross sections. Both models showed good fit to the data observados with a determination coefficient of 0.977 for the model obtained from multiple linear regression and to 0.99 obtained from dimensional analysis. The model obtained from the dimensional analysis considers the change in diameter than the tube suffers the water pressure. Both models are easy to use.

Introdução - A maior limitação para avaliar a dieta habitual é dada pelo erro de medida. Para minimizar os seus efeitos, tem-se proposto metodologias de calibração para correção dos dados e medidas associativas, consistindo na determinação de uma relação entre duas escalas de medida, utilizando-se regressão linear. Objetivo - Calibrar os dados obtidos por Questionário de Freqüência Alimentar para Adolescentes (QFAA), a partir do fator de calibração obtido por regressão linear. Metodologia - A amostra foi constituída por 74 adolescentes de ambos os sexos, alunos de uma escola pública de Piracicaba, com idade entre 10 e 14 anos. Foram excluídos indivíduos com idade superior ou igual a 14 anos ou com dados de consumo de energia não plausíveis. Obtiveram-se informações sobre dados socioeconômicos, antropométricos, demográficos e de maturação sexual. O consumo alimentar foi levantado por meio de Questionário de Freqüência Alimentar para Adolescentes e dois Recordatórios de 24 horas. Os dados de consumo de ambos os instrumentos foram ajustados pela energia, sendo que apenas os dados do recordatório foram ajustados pela variabilidade intrapessoal. Realizaram-se análises descritivas e de tendência central, one way ANOVA, coeficientes de correlação de Pearson e regressão linear. A média dos dois recordatórios foi utilizada como referência para a calibração dos dados. Resultados - De 74 indivíduos, 71,6% eram do sexo feminino. As médias das variáveis dietéticas foram muito semelhantes para o questionário calibrado e média dos recordatórios, com redução dos valores de desvio-padrão. Os coeficientes de calibração da regressão linear variaram de -0,05 (ferro) a 0,28 (vitamina C). Conclusão – Pelos resultados encontrados, a metodologia utilizada para a calibração dos dados dietéticos foi capaz de reduzir o erro de mensuração e, mesmo não o eliminando por completo, é uma abordagem que pode ser utilizada para obter estimativas menos enviesadas. / Background - A major limitation in usual diet assessment is the measurement error. Calibration approaches have been proposed to minimize its effects and to correct risk estimates. Calibration could be defined as a method which determines a relation between rank orders of two instruments by linear regression. Objective – to apply a calibration strategy in nutrient intake datas obtained by Adolescents Food Frequency Questionnaire (AFFQ), by using a calibration factor obtained by linear regression. Methodology – 74 boys and girls (10 to 14 years old) enrolled at a public school of Piracicaba were assessed. Values of energy intake higher than 6000Kcal and adolescents older than 14 years were excluded. Demographic and anthropometric data, sexual maturation and dietary intake (assessed by food frequency questionnaire and 24-hour recall) were examined. Dietary data intakes were adjusted by energy and, only 24-hour recall data were adjusted by within-person variance. Descriptive statistics, one way ANOVA, Pearson correlation coefficients and linear regression were performed. Results - 71,6% were girls. The calibrated values were similar to the reference data, with a reduction of standard deviation values. Linear regression coefficients (λ) ranged from -0,05 (iron) to 0,28 (vitamin C). Conclusion – The methodology used to calibrate dietary data was capable to reduce measurement error. Although it was not able to eliminate error completely, it is an approach that can be used to obtain less unbiased estimates.

This master thesis deals with the analysis of housing prices in Prague. The main goal is to identify and explain the factors which have an influence on the prices of flats at the macro and micro level. Two spatial statistic methods, namely multiple linear regressions and geographically weighted regressions (GWR), are used in the first part of the thesis, which deals with the prices in Prague in general. The influence on the values of flats in Prague basic settlement units caused by several factors such as the distance from the Old Town Square, age of dwellings, the presence of migrants or air pollution was investigated using these two methods. The price map of the association of real estate agencies, the Czech Statistical Office and the Prague Institute of Planning and Development provided the data used in the presented research. Price profiles from the centre of Prague to the suburbs in various directions were also created and analyzed. Factors with an influence on housing prices at the micro level in a case study of the Prague cadastral territory of Modřany are described in the second part of the thesis. The analysis of new developer projects and older flats in panel houses investigates the influence on the housing prices caused by factors such as noise, physical condition of apartments and the quality of...

Approved for public release; distribution is unlimited. / This paper documents the development of a statistical analysis package for the TRS-80 microcoraputer. The package is comprised of six interactive programs which are generally divided into topical areas. The major emphasis is on exploratory data analysis and statistical inference, however, probability and inverse probability distributions are also included. The programming language is TRS-80 Level II BASIC enhanced by the input/output commands available through the ESF-80 (Exatron Stringy Floppy) mass storage subsystem. With the modification of these few commands, the package is compatible with most floppy disk operating systems designed for the TRS-80 Model I or Model III microcomputers. This statistical analysis capability implemented on a relatively inexpensive system provides a useful tool to the student or the trained analyst without ready access to a mainframe computer system. / Major, United States Marine Corps

The diploma thesis describes migration and migration policy development of the Czech and Slovak republic, as well as theoretical studies dealing with migration effects on the labour market and on public finance of the migrants' destination countries. The core of the diploma thesis is an analysis of the relationship between macroeconomic indicators of the analysed countries and their migration flows, analysed by the regression analysis. The regression analysis is followed by a specific quantification of a potential impact of migration flows increase on public finance of the analysed countries. The main contribution of this diploma thesis is that it points out the connection between fiscal policy, migration flows and public finance of analysed countries. The main data sources are the OECD (2014) and EUROSTAT (2014) database and the analysed period is the year 1998 to 2012. I have come to the conclusion that there is a linear relationship between the increase of the selected macroeconomic indicators and migration flows of the analysed countries in case of the Czech and Slovak republic. This increase has negative or positive impact on their public finance, depending on type of increased macroeconomic indicator.

No description available.

This master‘s thesis deals with impacts on prices of sites which are meant by planning for residential housing in the landscape park Moravian Karst. In the research section of the thesis, there are defined fundamental concepts that concern themselves with the given issues and there are portrayed impacts on prices of sites according to the accessible literature. In the analytical section of the thesis, there is conducted an analysis of the real estate market of the given segment. Furthermore, there are examined and statistically evaluated impacts on prices of sites based on a multiple linear regression modelled in a statistical tool gretl. Impacts are evaluated both within a scope of the whole area and within particular municipalities. In the discussion section of the thesis, the manifested questions are answered, the stated hypotheses are tested and the obtained results are compared to the findings that were described in the research section.

The goal of this thesis is introduce the theory of generalized linear models, namely probit and logit model. This models are especially used for medical data processing. In our concrete case these mentioned models are applied to data file obtained in teaching hospital Brno. The aim is statically analyzed immune response of child patients in dependence of twelve selected types of genes and find out which combinations of these genes influence septic state of patients.

In the iron ore mining fraternity, in order to achieve the desired quality in the froth flotation processing plant, stakeholders rely on conventional laboratory test technique which usually takes more than two hours to ascertain the two variables of interest. Such a substantial dead time makes it difficult to put the inherent stochastic nature of the plant system in steady-state. Thus, the present study aims to evaluate the feasibility of using machine learning algorithms to predict the percentage of silica concentrate (SiO2) in the froth flotation processing plant in real-time. The predictive model has been constructed using iron ore mining froth flotation system dataset obtain from Kaggle. Different feature selection methods including Random Forest and backward elimination technique were applied to the dataset to extract significant features. The selected features were then used in Multiple Linear Regression, Random Forest and Artificial Neural Network models and the prediction accuracy of all the models have been evaluated and compared with each other. The results show that Artificial Neural Network has the ability to generalize better and predictions were off by 0.38% mean square error (mse) on average, which is significant considering that the SiO2 range from 0.77%- 5.53% -( mse 1.1%) . These results have been obtained within real-time processing of 12s in the worst case scenario on an Inter i7 hardware. The experimental results also suggest that reagents variables have the most significant influence in SiO2 prediction and less important variable is the Flotation Column.02.air.Flow. The experiments results have also indicated a promising prospect for both the Multiple Linear Regression and Random Forest models in the field of SiO2 prediction in iron ore mining froth flotation system in general. Meanwhile, this study provides management, metallurgists and operators with a better choice for SiO2 prediction in real-time per the accuracy demand as opposed to the long dead time laboratory test analysis causing incessant loss of iron ore discharged to tailings.