In petroleum production systems, hydrate morphology is observed to be influenced by the crude oil composition. This work is aimed at identifying which crude oil compositional parameters that need to be determined in order to evaluate natural anti-agglomerating properties of crude oils, i.e. the critical compositional descriptors. The compositional features of 22 crude oils have been studied, and multivariate data analysis has been used to investigate the possibility for correlations between several crude oil properties. The results show that biodegradation together with a relatively large amount of acids are characteristic for non-plugging crude oils, while excess of basic compounds is characteristic for plugging crude oils. The multivariate data analysis shows a division of the nonbiodegraded oils, which are all plugging, and the biodegraded oils. In addition, the biodegraded oils seem to be divided into two groups, one with plugging oils and one with mostly non-plugging oils. The results show that the wettability can be predicted from the variables biodegradation level, density, asphaltene content and TAN.

O presente trabalho apresenta proposições para seleção de variáveis em avaliações sensoriais descritivas e de espectro infravermelho que contribuam com a indústria de alimentos e química através da utilização de métodos de análise multivariada. Desta forma, os objetivos desta tese são: (i) Estudar as principais técnicas de análise multivariada de dados, como são comumente organizadas e como podem contribuir no processo de seleção de variáveis; (ii) Identificar e estruturar técnicas de análise multivariada de dados de forma a construir um método que reduza o número de variáveis necessárias para fins de caracterização, classificação e predição dos produtos; (iii) Reduzir a lista de variáveis/atributos, selecionando aqueles relevantes e não redundantes, reduzindo o tempo de execução e a fadiga imposta aos membros de um painel em avaliações sensoriais; (iv) Validar o método proposto utilizando dados reais; e (v) Comparar diferentes abordagens de análise sensorial voltadas ao desenvolvimento de novos produtos. Os métodos desenvolvidos foram avaliados através da aplicação de estudos de caso, em exemplos com dados reais. Os métodos sugeridos variam com as características dos dados analisados, dados altamente multicolineares ou não e, com e sem variável dependente (variável de resposta). Os métodos apresentam bom desempenho, conduzindo a uma redução significativa no número de variáveis e apresentando índices de adequação de ajuste dos modelos ou acurácia satisfatórios quando comparados aos obtidos mediante retenção da totalidade das variáveis ou comparados a outros métodos dispostos na literatura. Conclui-se que os métodos propostos são adequados para a seleção de variáveis sensoriais e de espectro infravermelho. / This dissertation presents propositions for variable selection in data from descriptive sensory evaluations and near-infrared (NIR) spectrum analyses, based on multivariate analysis methods. There are five objectives here: (i) review the main multivariate analysis techniques, their relationships and potential use in variable selection procedures; (ii) propose a variable selection method based on the techniques in (i) that allows product prediction, classification, and description; (iii) reduce the list of variables/attributes to be analyzed in sensory panels identifying those relevant and non-redundant, such that the time to collect panel data and the fatigue imposed on panelists is minimized; (iv) validate methodological propositions using real life data; and (v) compare different sensory analysis approaches used in new product development. Proposed methods were evaluated through case studies, and vary according to characteristics in the datasets analyzed (data with different degrees of multicollinearity, presenting or not dependent variables). All methods presented good performance leading to significant reduction in the number of variables in the datasets, and leading to models with better adequacy of fit. We conclude that the methods are suitable for datasets from descriptive sensory evaluations and NIR analyses.

O presente trabalho apresenta proposições para seleção de variáveis em avaliações sensoriais descritivas e de espectro infravermelho que contribuam com a indústria de alimentos e química através da utilização de métodos de análise multivariada. Desta forma, os objetivos desta tese são: (i) Estudar as principais técnicas de análise multivariada de dados, como são comumente organizadas e como podem contribuir no processo de seleção de variáveis; (ii) Identificar e estruturar técnicas de análise multivariada de dados de forma a construir um método que reduza o número de variáveis necessárias para fins de caracterização, classificação e predição dos produtos; (iii) Reduzir a lista de variáveis/atributos, selecionando aqueles relevantes e não redundantes, reduzindo o tempo de execução e a fadiga imposta aos membros de um painel em avaliações sensoriais; (iv) Validar o método proposto utilizando dados reais; e (v) Comparar diferentes abordagens de análise sensorial voltadas ao desenvolvimento de novos produtos. Os métodos desenvolvidos foram avaliados através da aplicação de estudos de caso, em exemplos com dados reais. Os métodos sugeridos variam com as características dos dados analisados, dados altamente multicolineares ou não e, com e sem variável dependente (variável de resposta). Os métodos apresentam bom desempenho, conduzindo a uma redução significativa no número de variáveis e apresentando índices de adequação de ajuste dos modelos ou acurácia satisfatórios quando comparados aos obtidos mediante retenção da totalidade das variáveis ou comparados a outros métodos dispostos na literatura. Conclui-se que os métodos propostos são adequados para a seleção de variáveis sensoriais e de espectro infravermelho. / This dissertation presents propositions for variable selection in data from descriptive sensory evaluations and near-infrared (NIR) spectrum analyses, based on multivariate analysis methods. There are five objectives here: (i) review the main multivariate analysis techniques, their relationships and potential use in variable selection procedures; (ii) propose a variable selection method based on the techniques in (i) that allows product prediction, classification, and description; (iii) reduce the list of variables/attributes to be analyzed in sensory panels identifying those relevant and non-redundant, such that the time to collect panel data and the fatigue imposed on panelists is minimized; (iv) validate methodological propositions using real life data; and (v) compare different sensory analysis approaches used in new product development. Proposed methods were evaluated through case studies, and vary according to characteristics in the datasets analyzed (data with different degrees of multicollinearity, presenting or not dependent variables). All methods presented good performance leading to significant reduction in the number of variables in the datasets, and leading to models with better adequacy of fit. We conclude that the methods are suitable for datasets from descriptive sensory evaluations and NIR analyses.

A função da contabilidade tem sido avaliar a riqueza do homem, avaliar os acréscimos ou decréscimos dessa riqueza. Avaliar a riqueza das organizações requer alguma medida, é necessário uma forma de mensurar desempenho de empresas. Portanto, é necessário sabermos qual é o reflexo, no âmbito acadêmico, destas medidas de desempenho, quantas e quais são as medidas de desempenho preferencialmente utilizadas pelos artigos científicos na área de contabilidade. E, ainda, é necessário observarmos como essas medidas se agrupam. Para isso o presente trabalho analisou pesquisas na área contábil que estudaram medidas de desempenho de empresas e identificou como as medidas de desempenho utilizadas nas pesquisas podem ser agrupadas pela técnica de análise fatorial. Como principal descoberta pode-se citar as inúmeras variáveis diferentes (237) encontradas na literatura contábil como proxy da mensuração de desempenho. A maioria (62,44%) das variáveis aparece apenas uma vez no universo de todos os trabalhos pesquisados. Isso evidencia a frequente criação de medidas de desempenho pelos autores. Tal fato se deve possivelmente a aplicações específicas de análise de desempenho, com certeza essas variáveis criadas são importantes para o contexto individual de cada pesquisa, porém a análise fatorial empregada por este trabalho demonstrou que de uma maneira geral, as variáveis sempre vão se agrupar conforme os indicadores da análise de balanço e as variáveis contábeis e pode-se perceber que não é necessário usarmos diversas variáveis, as mais frequentes conseguem retratar quase a totalidade (mais de 98% no ano de 2003) das variáveis originais de um modelo. E mais, as medidas com maior peso de explicação da variância dos dados originais são aquelas variáveis extraídas dos demonstrativos contábeis, ou seja, um apoio à afirmativa de que a Contabilidade é fundamental para uma melhor tomada de decisão na área empresarial. / In this work, measurements for companies performance present in the accounting literature are analyzed. Performance measurements are important in order to inform the accounting users about the company status. On the other hand, we have performance measurements been used in the researches and, an interesting topic is to investigate, in the academic literature their frequency of occurrence, or how these measurements can be arranged in groups. To do this, articles from journals and conferences, as well as academic thesis in Brazil, were analyzed. In addition, a factorial analysis was done in order to group of measurements into factors. As a result it was found many different variables (237 performance measurements) present in the accounting researches. The 62.44% of variables appears only once in the selected sample. This fact should possibly come from specific applications of performance analysis, where such measurements are important. However, the factorial analysis performed to a specific sector, has confirmed that, the measurements form factors, according to the knowing groups. As inferred by the factorial analysis, there is no need to use multiple variables; they can be grouped in traditional groups. The most relevant measurements, explaining the variance of the original data, are those extracted directly from accounting statements, reinforcing that accounting is essential for better decision making.

Taxonomia vegetal é a ciência que trata da descrição, identificação, nomenclatura e classificação de plantas. O desenvolvimento de novas técnicas que podem ser aplicadas nesta área de conhecimento é essencial para dar suporte às decisões relacionadas a conservação de hotspots de biodiversidade. Nesta dissertação de mestrado foi desenvolvido um protocolo de metabolic fingerprinting utilizando MALDI-MS (matrix-assisted laser desorption/ionisation mass spectrometry) e subsequente análise multivariada utilizando scripts desenvolvidos para o pacote estatístico R. Foram classificadas, com base nos seus metabólitos detectados, 24 plantas de diferentes famílias vegetais, sendo todas elas coletadas em áreas da Savana Brasileira (Cerrado), que foi considerada um hotspot de biodiversidade. Metabolic fingerprinting compreende uma parte da Metabolômica, i.e., a ciência que objetiva analisar todos os metabólitos de um dado sistema (celula, tecído ou organismo) em uma dada condição. Comparada com outros métodos de estudo do metaboloma MALDI-MS apresenta a vantagem do rápido tempo de análise. A complexidade e importância da correta classificação taxonômica é ilustrada no exemplo do gênero Lychnophora, o qual teve diversas espécies incluídas neste estudo. No Brasil espécies deste gênero são popularmente conhecidas como \"arnica da serra\" ou \"falsa arnica\". Os resultados obtidos apontam similaridades entre a classificação proposta e a classificação taxonômica atual. No entanto ainda existe um longo caminho para que a técnica de metabolic fingerprinting possa ser utilizada como um procedimento padrão em taxonomia. Foram estudados e discutidos diversos fatores que afetaram os resultados como o preparo da amostra, as condições de análise por MALDI-MS e a análise de dados, os quais podem guiar futuros estudos nesta área de pesquisa. / Plant taxonomy is the science of description, identification, nomenclature and classification of plants. The development of new techniques that can be applied in this field of research are essential in order to assist informed and efficient decision-making about conservation of biodiversity hotspots. In this master\'s thesis a protocol for metabolic fingerprinting by matrix-assisted laser desorption/ionisation mass spectrometry (MALDI-MS) with subsequent multivariate data analysis by in-house algorithms in the R environment for the classification of 24 plant species from closely as well as from distantly related families and tribes was developed. Metabolic fingerprinting forms part of metabolomics, a research field, which aims to analyse all metabolites, i.e., the metabolome in a given system (cell, tissue, or organism) under a given set of conditions. Compared to other metabolomics techniques MALDI-MS shows potential advantages, mainly due to its rapid data acquisition. All analysed species were collected in areas of the Brazilian Savanna (Cerrado), which was classified as \"hotspot for conservation priority\". The complexity and importance of correct taxonomic classification is illustrated on the example of the genus Lychnophora, of which several species also have been included into analysis. In Brazil species of this genus are popularly known as \"arnica da serra\" or \"falsa arnica\". Similarities to taxonomic classification could be obtained by the proposed protocol and data analysis. However there is still a long way to go in making metabolic fingerprinting by MALDI-MS a standard procedure in taxonomic research. Several difficulties that are inherent to sample preparation, analysis of plant\'s metabolomes by MALDI-MS as well as data analysis are highlighted in this study and might serve as a basis for further research.

Orientadores: Juliana Azevedo Lima Pallone, Ronei Jesus Poppi / Tese (doutorado) - Universidade Estadual de Campinas, Faculdade de Engenharia de Alimentos / Made available in DSpace on 2018-08-22T02:38:43Z (GMT). No. of bitstreams: 1 Ferreira_DanielaSouza_D.pdf: 2148960 bytes, checksum: 45b2c46dd9d82dfd00f454c5a430d494 (MD5) Previous issue date: 2013 / Resumo: A avaliação da qualidade nutricional de alimentos é realizada principalmente por meio da determinação dos componentes majoritários, conhecida como composição centesimal (umidade, proteína, cinza, lipídio, carboidrato e fibra). No entanto, os métodos tradicionais de análise são demorados e utilizam materiais, equipamentos e diversos reagentes químicos, que além de oferecerem risco ao analista, geram resíduos tóxicos. Diante disto, uma alternativa para a análise química de grãos, rápida, de baixo custo e sem uso de reagentes químicos é a espectroscopia na região do infravermelho. Visando atender a demanda do Brasil por pesquisas empregando espectroscopia no infravermelho para análise de alimentos, o objetivo desse trabalho foi avaliar a possibilidade de utilização das técnicas espectroscopia no infravermelho próximo (NIR), principalmente, e médio MIR, associadas à quimiometria, para previsão de parâmetros de qualidade da soja brasileira e quinoa da América do Sul. Para comparar a aplicação de NIR e MIR, amostras de soja provenientes do Paraná foram analisadas pelas duas técnicas para previsão da composição centesimal. Os erros relativos (E%) entre os valores de referência e os valores previstos pelos modelos de calibração PLS, foram pequenos tanto para o NIR como para o MIR, no entanto, os resultados sugerem o uso de NIR para previsão de lipídios (0,2 a 9,2%) e o uso de MIR para proteínas (0,2 a 5,6%), cinzas (0 a 5,0%) e umidade (0,1 a 2,0%). Posteriormente, foram construídos modelos de calibração PLS com NIR para previsão dos parâmetros de qualidade em soja moída e para a quinoa, grão inteiro e moído. Os melhores modelos de calibração para soja encontrados neste estudo foram para o conteúdo de proteína e umidade, com melhores coeficientes de determinação e raiz quadrada do erro médio quadrático de calibração (R2= 0,81, RMSEC = 0,58% e R2 = 0,80, RMSEC = 0,28%, respectivamente), contudo, a técnica mostrou capacidade adequada de predição para todos os parâmetros, incluindo lipídios, cinzas, carboidratos e fibras. Para amostras de quinoa, os espectros NIR foram inicialmente submetidos a uma análise de componentes principais (PCA) para tentar separá-las em grupos, de acordo com a origem geográfica destes grãos, os quais eram provenientes do Brasil, Bolívia e Peru. Duas componentes principais explicaram 98,2% do total da variância e três grupos foram observados na separação por PCA de acordo com o país de origem. A técnica de calibração por PLS produziu modelos adequados, que permitiu a quantificação da composição majoritária tanto para o grão inteiro como farinha de quinoa, mostrando boa correlação entre o valor previsto e o valor real, com R2 > 0,65 e RMSEC< 1,70%. Portanto, este estudo demonstra que a técnica de NIR é potencialmente útil como um método analítico não destrutivo para determinações rápidas e simples de constituintes alimentares, além de não necessitar nenhum tipo de preparo de amostra, já que os espectros dos grãos inteiros de quinoa forneceram bons resultados para previsão dos parâmetros estudados / Abstract: Evaluation of nutritional quality of food has been mainly performed by determination of major compounds, which is known as centesimal composition (moisture, protein, ash, lipid, carbohydrate and fiber). However, the traditional methods of analysis are time-consuming, use many materials and equipment, and also toxic reagents, that generate waste and are a risk for the analyst. Thus, infrared spectroscopy is an alternative to chemical analysis of grains, as it is a rapid, low cost technique and it does not use toxic reagents. In coming years, Brazilian researches using infrared for food analysis should increase, thus the objective of this work was to evaluate the possibility of application mainly of near-infrared (NIR) and mid-infrared (MIR) spectroscopy techniques coupled with chemometrics to predict quality parameters in Brazilian soybean and South America quinoa. In order to compare NIR and MIR techniques, the soybean group from Paraná (Brazil) was analyzed using both techniques to predict centesimal composition. The related errors (E%) between reference values and predicted values by partial least square (PLS) were low for both the NIR and the MIR. However, the results propose the use of NIR to predict lipid (E% of 0.2 to 9.2) content and the use of MIR to predict protein (E% of 0.2 to 5.6), ash (E% of 0 to 5.0), and moisture (E% of 0.1 to 2.0) contents. Subsequently, PLS regression models were constructed using NIR to predict quality parameters in ground soybean and quinoa, grain and ground. The best calibration models to soybean found in this study were the ones used to determine protein and moisture content (R2 = 0.81, RMSEP = 1.61% and R2 = 0.80, RMSEC = 1.55%, respectively). However, the technique shows high predictability for all parameters, including lipids, ash, carbohydrates and fibers, RMSECV of 0.40 to 2.30% and RMSEP 0.38 to 3.71%. For quinoa samples NIR spectra were obtained and principal component analysis (PCA) was applied to try to identify the geographic origin of quinoa samples, from Brazil, Peru and Bolivia. Two principal components explained 98.3% of the total variance and three groups were observed using PCA. The PLS models developed for the chemical composition showed that the proposed methodology produced adequate results, as whole grain as ground quinoa, with the graph of the real and predicted concentration having a coefficient of determination (R2) > 0.65 and RMSEC < 1.70%. The viability of the NIR technique with no waste generation, low cost, reduced time and no kind of sample preparation for replacing laborious methods of analysis was demonstrated because the results for grains were satisfactory / Doutorado / Ciência de Alimentos / Doutora em Ciência de Alimentos

Enteric (or typhoid) fever is a systemic infection mainly caused by Salmonella Typhi and Salmonella Paratyphi A. The disease is common in areas with poor water quality and insufficient sanitation. Humans are the only reservoir for transmission of the disease. The presence of asymptomatic chronic carriers is a complicating factor for the transmission. There are major limitations regarding the current diagnostic methods both for acute infection and chronic carriage. Metabolomics is a methodology studying metabolites in biological systems under influence of environmental or physiological perturbations. It has been applied to study several infectious diseases, with the goal of detecting diagnostic biomarkers. In this thesis, a mass spectrometry-based metabolomics approach, including chemometric bioinformatics techniques for data analysis, has been used to evaluate the potential of metabolite biomarker patterns for diagnosis of enteric fever at different stages of the disease. In Paper I, metabolite patterns related to acute enteric fever were investigated. Human plasma samples from patients in Nepal with culture-confirmed S. Typhi or S. Paratyphi A infection were compared to afebrile controls. A metabolite pattern discriminating between acute enteric fever and afebrile controls, as well as between the two causative agents of enteric fever was detected. The strength of using a panel of metabolites instead of single metabolites as biomarkers was also highlighted. In Paper II, metabolite patterns for acute enteric fever, this time focusing only on S. Typhi infections, were investigated. Human plasma from patients in Bangladesh with culture-positive or -negative but clinically suspected S. Typhi infection were compared to febrile controls. Differences were found in metabolite patterns between the culture-positive S. Typhi group and the febrile controls with a heterogeneity among the suspected S. Typhi samples. Consistencies in metabolite patterns were found to the results from Paper I. In addition, a validation cohort with culture-positive S. Typhi samples and a control group including patients with malaria and infections caused by other pathogens was analysed. Differences in metabolite patterns were detected between S. Typhi samples and all controls as well as between S. Typhi and malaria. Consistencies in metabolite patterns were found to the primary Bangladeshi cohort and the Nepali cohort from Paper I. Paper III focused on chronic Salmonella carriers. Human plasma samples from patients in Nepal undergoing cholecystectomy with confirmed S. Typhi or S. Paratyphi A gallbladder carriage were compared to non-carriage controls. The Salmonella carriage samples were distinguished from the non-carriage controls and differential signatures were also found between the S. Typhi and S. Paratyphi A carriage samples. Comparing metabolites found during chronic carriage and acute enteric fever (in Paper I) resulted in a panel of metabolites significant only during chronic carriage. This work has contributed to highlight the potential of using metabolomics as a tool to find diagnostic biomarker patterns associated with different stages of enteric fever.

The continuing need for greater energy security and energy independence has motivated researchers to develop new energy technologies for better energy resource management and efficient energy usage. The focus of this dissertation is the development of optical (spectroscopic) sensing methodologies for various fuels, and energy applications. A fiber-optic NIR sensing methodology was developed for predicting water content in bio-oil. The feasibility of using the designed near infrared (NIR) system for estimating water content in bio-oil was tested by applying multivariate analysis to NIR spectral data. The calibration results demonstrated that the spectral information can successfully predict the bio-oil water content (from 16% to 36%). The effect of ultraviolet (UV) light on the chemical stability of bio-oil was studied by employing laser-induced fluorescence (LIF) spectroscopy. To simulate the UV light exposure, a laser in the UV region (325 nm) was employed for bio-oil excitation. The LIF, as a signature of chemical change, was recorded from bio-oil. From this study, it was concluded that phenols present in the bio-oil show chemical instability, when exposed to UV light. A laser-induced breakdown spectroscopy (LIBS)-based optical sensor was designed, developed, and tested for detection of four important trace impurities in rocket fuel (hydrogen). The sensor can simultaneously measure the concentrations of nitrogen, argon, oxygen, and helium in hydrogen from storage tanks and supply lines. The sensor had estimated lower detection limits of 80 ppm for nitrogen, 97 ppm for argon, 10 ppm for oxygen, and 25 ppm for helium. A chemiluminescence-based spectroscopic diagnostics were performed to measure equivalence ratios in methane-air premixed flames. A partial least-squares regression (PLS-R)-based multivariate sensing methodology was investigated. It was found that the equivalence ratios predicted with the PLS-R-based multivariate calibration model matched with the experimentally measured equivalence ratios within 7 %. A comparative study was performed for equivalence ratios measurement in atmospheric premixed methane-air flames with ungated LIBS and chemiluminescence spectroscopy. It was reported that LIBS-based calibration, which carries spectroscopic information from a “point-like-volume,” provides better predictions of equivalence ratios compared to chemiluminescence-based calibration, which is essentially a “line-of-sight” measurement.

No description available.

Water quality monitoring is crucial globally due to the vital role of freshwater in providing drinking water, irrigation, and ecosystem services. Highly polluted water poses risks to both ecosystems and human health. Current water quality monitoring methods deployed in the field are often expensive, labor-intensive, and invasive. To overcome these issues, this degree project investigated the use of remote sensing to assess critical water quality parameters in the Swedish river Indalsälven. The research questions focus on determining the accuracy of predicting chemical oxygen demand (COD), river color, turbidity, and total phosphorus (TP) using satellite data and machine learning algorithms. The findings revealed that COD can be predicted with a cross-validated coefficient of determination (R²CV) of 0.7, indicating a robust predictive capability. The study suggests that while approximate quantitative prediction of COD in oligotrophic rivers is feasible using Sentinel-2 imagery, predictions for the other parameters remain challenging in the context of Indalsälven. Improvements in prediction accuracy were achieved through optimized band combinations, reduced datasets encompassing satellite data collected within two days of field measurements, and suitable pre-processing methods. / Airborne Monitoring of Water Quality in Remote Regions