1 |
Aplicação de métodos de resolução de curvas multivariada a dados experimentais gerados por diferentes técnicas analíticas = uma visão exploratória da biossíntese da violaceína / Application of multivariate curve resolution of experimental data generated by several analytical techniques : an exploratory overview of violacein biosynthesisMuniz Filho, Raimundo Clecio Dantas 07 November 2011 (has links)
Orientador: Márcia Miguel Castro Ferreira / Tese (doutorado) - Universidade Estadual de Campinas, Instituto de Quimica / Made available in DSpace on 2018-08-19T01:35:46Z (GMT). No. of bitstreams: 1
MunizFilho_RaimundoClecioDantas_D.pdf: 4450185 bytes, checksum: 93734fb6cdec46f7ac2161a230dd978c (MD5)
Previous issue date: 2011 / Resumo: A aplicação de métodos de resolução de curvas multivariada (MCR-ALS e PARAFAC) a dados de espectroscopia de fluorescência multicomprimento de onda, absorção molecular e cromatografia líquida de ultra eficiência com detecção por espectrometria de massas foi avaliada como um nova e complementar metodologia para se obter informação inicial sobre a biossíntese de produtos naturais. Para demonstrar o poder dessa proposta, a biossíntese da violaceína pela Chromobacterium violaceum foi usada como exemplo. A violaceína é um metabólito secundário que exibe amplo espectro de atividade biológica e o entendimento de sua rota biossintética pode ser de grande interesse para o planejamento de fármacos. Para tanto, uma cultura da C. violaceum foi cultivada em biorreator, onde, a cada duas horas, alíquotas do meio de cultura eram coletadas para subsequentes análises. Na primeira parte deste trabalho, as matrizes de emissão-excitação das amostras revelaram o comportamento dinâmico de fluoróforos envolvidos na biossíntese. O método de análise de fatores paralelos (PARAFAC) foi aplicado para resolver os perfis de fluorescência, onde seis espectros puros de emissão e excitação foram obtidos e suas relações com a biossíntese da violaceína estabelecidas. Na segunda parte, o método de resolução de curvas multivariada por quadrados mínimos alternados (MCR-ALS) foi utilizado para extrair informações simultâneas dos dados gerados por UV-VIS, fluorescência e UPLC-MS das amostras coletadas durante a biossíntese. Cinco compostos foram resolvidos e informações espectrais e temporais dessas espécies foram obtidas. Compostos com massas moleculares de 453, 465 e 479 u, até então não reportados na literatura como intermediários dessa biossíntese, foram identificados. Por fim, a metodologia baseada na aplicação de métodos de resolução de curvas MCR-ALS e PARAFAC provou ser uma poderosa ferramenta para investigar a biossíntese de produtos naturais / Abstract:The application of Multivariate Curve Resolution methods (MCR-ALS and PARAFAC) into multi-wavelength fluorescence spectroscopy, molecular absorption and ultra-performance liquid chromatography coupled to mass spectrometry data was evaluated as a new and complementary approach for the initial analysis of the biosynthesis of fluorescent natural products. To demonstrate the usefulness of this methodology, violacein biosynthesis by Chromobacterium violaceum was used as an example. Violacein is a secondary metabolite that exhibits a broad spectrum of biological activities and the understanding of its biosynthetic pathway can be of great interest for drug design. Therefore, a culture of C. violaceum was grown in a bioreactor from which aliquots were collected every two hours for subsequent analysis. In the first part of this work, the excitation¿emission matrices of the samples have revealed the dynamic behavior of the fluorophores involved in this biosynthesis. The parallel factor analysis (PARAFAC) method was applied to resolve the fluorescence profiles where six pure emission and excitation spectra were obtained and their relationship with violacein biosynthesis established. In the second part, the multivariate curve resolution procedure based on alternating least squares optimization (MCR-ALS) was used for extracting information from fused data (UV-VIS absorption, fluorescence and ultra-performance liquid chromatography with mass spectrometry) gathered during the violacein biosynthesis. Five chemical components were resolved from which spectral and temporal information were obtained. Compounds with molecular masses 453, 465 and 479 u, until now unreported as intermediate of this biosynthesis, were identified. Finally, the methodology based on the application of curve resolution methods PARAFAC and MCR-ALS has proven to be a unique and powerful tool to achieve a deeper insight of the biosynthesis of natural products / Doutorado / Físico-Química / Doutor em Ciências
|
2 |
Desenvolvimento do metodo multivariado acelerado para determinação do prazo de validade de produtos unindo quimiometria e cinetica quimica / Development of the multivariate accelerated shelf-life test for determining the shelf-life of products using chemometrics and chemical kineticsPedro, Andre Messias Krell 14 August 2018 (has links)
Orientador: Marcia Miguel Castro Ferreira / Tese (doutorado) - Universidade Estadual de Campinas, Instituto de Quimica / Made available in DSpace on 2018-08-14T13:05:23Z (GMT). No. of bitstreams: 1
Pedro_AndreMessiasKrell_D.pdf: 1179705 bytes, checksum: 302be0a6ab2e796d557a625c76f92bc4 (MD5)
Previous issue date: 2009 / Resumo: DESENVOLVIMENTO DO MÉTODO MULTIVARIADO ACELERADO PARA DETERMINAÇÃO DO PRAZO DE VALIDADE DE PRODUTOS UNINDO QUIMIOMETRIA E CINÉTICA QUÍMICA descreve um novo conceito de análise de dados que permite avaliar os mecanismos que governam a degradação de bens de consumo e determinar o período de tempo no qual os bens de consumo mantém suas características dentro de níveis aceitáveis. O algoritmo une técnicas quimiométricas com a teoria cinética química formal, constituindo um avanço sobre os protocolos para determinação da validade de produto existentes. Além de ser de fácil interpretação, suas principais vantagens incluem a capacidade de unir informações provenientes de diferentes disciplinas ¿ analíticos, físico-químicos, sensoriais - e a possibilidade de utilização direta de dados oriundos de instrumentação analítica como espectroscópios, cromatógrafos, calorímetros, etc. O trabalho é dividido em seis capítulos. O Capítulo I traz uma revisão da cinética química e dos métodos convencionais para determinação do prazo de validade de produtos. O algoritmo do Método Multivariado Acelerado, bem como suas premissas, vantagens e desvantagens, são descritos no capítulo II. Os Capítulos III e IV trazem aplicações do Método Multivariado Acelerado. No Capítulo III dados físico-químicos e sensoriais foram utilizados para determinar o prazo de validade de um produto alimentíco. No Capítulo IV utilizaram-se dados instrumentais, oriundos de espectroscopia no infravermelho próximo (NIR), como fonte de informação quantitativa que, aliados a avaliações sensoriais, permitiram a determinação da validade de um produto cosmético e forneceram informações relevantes sobre seus modos de degradação. Uma comparação com o método tri-linear PARAFAC é apresentada no Capítulo V. Apesar de constituir ferramenta importante para avaliação dos modos de degradação de produtos, o método PARAFAC carece de estratégias para determinação efetiva do prazo de bens de consumo. Finalmente, o Capítulo VI mostra o desenvolvimento de um protocolo para determinação de várias propriedades de produtos concentrados de tomate utilizando regressão PLS2. Demonstrou-se que, apesar da preferência pelo método de calibração PLS1 por parte dos pesquisadores, a regressão PLS2 é vantajosa quando há correlação entre propriedades cujos métodos de referência são precisos e outras, cuja quantificação é imprecisa. / Abstract: DEVELOPMENT OF THE MULTIVARIATE ACCELERATED SHELF-LIFE TEST (MASLT) FOR DETERMINING THE SHELF-LIFE OF PRODUCTS BY MERGING CHEMOMETRICS AND CHEMICAL KINETICS describes a new concept of data analysis for evaluating the mechanisms that rule out the degradation of consumer goods. The algorithm herein developed took advantage of the soft modeling concepts from chemometrics and of chemical kinetics for determining the period of time within which products would be able to keep their quality characteristics within acceptable levels. Besides being of easy interpretation, the main advantages of the MASLT are the capacity to merge data from different sources ¿ analytical, physical chemical, sensorial ¿ and the ability of handling instrumental data (spectroscopic, chromatographic, calorimetric, etc.) directly for determining the shelf-life of products. This work and some applications are presented in six chapters. In Chapter I, a review of the chemical kinetics theory, as well as the concepts of conventional shelf-life methods, is presented. The MASLT algorithm, together with its assumptions, characteristics and advantages, is presented in Chapter II. Chapters III e IV describe applications where the shelf-life was determined successfully by using the MASLT. In Chapter III, physical chemical and sensorial data were merged not only for determining the shelf-life of an industrialized tomato product, but also to study the correlation between the various properties. In Chapter IV, instrumental data from NIR spectroscopy were used as source of quantitative information and, together with sensory analyses, allowed the determination of the shelf-life of a cosmetic product as well as provided valuable information with regards to its degradation mechanisms. A comparison between the tri-linear method PARAFAC is presented in Chapter V. Despite being an important tool for evaluating the degradation mechanisms of consumer goods, the PARAFAC method lacks strategies for effectively determining the shelf-life of products. Finally, Chapter VI describes the development of a new protocol for determining several properties of tomato concentrate products using PLS2 regression. It was demonstrated that, despite most researchers having a preference for using PLS1 regression, PLS2 is advantageous when there is strong correlations between properties that can be determined quite accurately by their reference methods with those presenting lower accuracy. / Doutorado / Físico-Química / Doutor em Ciências
|
3 |
Kinetic measurements using time-resolved NMR spectroscopyKhajeh, Maryam January 2010 (has links)
Many thousands of pounds are spent every year by pharmaceutical companies on understanding the mechanisms and kinetics of chemical reactions involved in drug discovery and production. NMR spectroscopy is often at the core of these studies as it is a powerful, non-destructive method for structure elucidation. As such investigations can be time-consuming and cost-inefficient, AstraZeneca, the project sponsor, is interested in more efficient methods for studying the kinetics of pharmaceutical reactions. In this work a number of different techniques have been devised, studied, and implemented to study the kinetics of chemical reactions by time-resolved NMR spectroscopy, in which every species in a reaction can be monitored simultaneously. These novel techniques allow the study of reactions which are difficult or impossible to study by conventional NMR methods (such as heterogeneous reactions), or which are complicated by having overlapping signals. It is possible to monitor the kinetics of a reaction very simply by acquiring a series of 1H spectra, and obtaining the integrals of the signals by least squares fitting. This technique has been used for kinetic studies of static and on-flow reactions. In the static systems the reaction mixture was placed in the normal NMR tube in the magnet, while in the flow system the reaction mixture was placed outside of the magnet, and the solution flowed through an NMR tube placed in the magnet. The novel flow system designed, constructed and tested here has been used for kinetic studies of illustrative homogeneous and heterogeneous reactions, and is suitable for use in a wide range of NMR instrumentation. Kinetic studies have also been carried out by acquiring a series of DOSY datasets, analysing the results using the multi-way method PARAFAC (PARAllel FACtor analysis). A series of DOSY datasets contains multivariate information on spectrum, time evolution and diffusion. Without providing any predetermined model, the data can be decomposed by PARAFAC to yield the spectrum, kinetics, and diffusion profiles for each of the components. It has also been shown that PARAFAC is remarkably robust to low signal-to-noise ratio data, significantly below the level at which conventional methods would fail.
|
4 |
Approches tensorielles pour les systèmes de communication MIMO avec relais / Tensor-based MIMO relaying communication systemsRonchini Ximenes, Leandro 25 March 2015 (has links)
Dans les communications coopératives, deux ou plusieurs terminaux de transmissionsont combinés pour accroître la diversité et/ou la puissance des signaux arrivant à un récepteur. Récemment, l'analyse tensorielle s'est avérée une approche efficace pour l'estimation de canaux dans les systèmes coopératifs. Cependant, parmi les quelques travaux consacrés à cette tâche, l'utilisation de la décomposition tensorielle PARAFAC pour modéliser les signaux reçus ne permet pas l'estimation conjointe des symboles et des canaux de communication. Afin d'éviter l'utilisation de séquences de symboles pilotes, l'objectif de cette thèse est de fournir de nouvelles approches tensorielles, en termes de systèmes de transmission et de récepteurs semi-aveugles, pour des systèmes de communication MIMO avec relai mono-directionnels, à deux sauts. Deux systèmes de transmission sont proposés en utilisant un codage spatio-temporel du type Khatri-Rao et deux stratégies de traitement Amplify-and-Forward (AF) au relai. Pour ces systèmes, appelés PT2-AF et NP-AF, les signaux reçus au niveau de la destination satisfont respectivement des modèles tensoriels du type PARATUCK2 et nested PARAFAC. En exploitant les propriétés d'unicité de ces modèles tensoriels établies dans la thèse, plusieurs récepteurs semi-aveugles sont dérivés. Certains de ces récepteurs sont du type ALS, tandis que d'autres sont des solutions non itératives basées sur des factorisations de produits de Khatri-Rao. Des résultats de simulation sont présentés pour illustrer les performances des récepteurs proposés qui sont comparés à des estimateurs supervisés. / In cooperative communication systems, two or more transmitting terminals arecombined to increase the diversity and/or the power of the signals arriving at aparticular receiver. Recently, the so-called tensor analysis has been an efficient approach for channel estimation in systems with cooperative diversity. However, among the few works devoted to this task, the utilization of the PARAFAC tensor decomposition for modeling the received signals did not allow the development of techniques for joint symbol and channel estimation. Aiming to avoid the use of pilot-based sequences, the objective of this thesis is to provide new tensor-based strategies, including transmission systems and semi-blind receivers, for one-way two-hop relaying systems. Based on a Khatri-Rao space-time coding at the source and two different Amplify-and-Forward (AF) relaying strategies, two transmission systems are proposed. For these systems, named PT2-AF and NP-AF, the received signals at the destination node follow respectively a PARATUCK2 and a nested PARAFAC tensor model. Exploiting uniqueness properties of these tensor models which are established in the thesis, several semi-blind receivers are derived. Some of these receivers are of iterative form using an ALS algorithm, whereas some other ones are close-form solutions associated with Khatri-Rao factorizations. Some simulation results are finally presented to illustrate the performance of the proposed receivers which are compared to some state-of-the-art supervised techniques.
|
5 |
Desenvolvimento de metodologia para a determinação de aflatoxinas em amostras de amendoim usando espectrofluorimetria e análise dos fatores paralelos (PARAFAC) / Development of methodology for the determination of aflatoxins in peanut samples using spectrofluorimetry and parallel factor analysisRobson Alves Luiz 21 March 2012 (has links)
Desenvolvimento de metodologia para a determinação de aflatoxinas em amostras de amendoim usando espectrofluorimetria e análise dos fatores paralelos (PARAFAC). 2012. 92 f. Dissertação (Mestrado em Engenharia Química) - Instituto de Química, Universidade do Estado do Rio de Janeiro , Rio de Janeiro, 2012. Neste trabalho de pesquisa são descritos dois estudos de caso que se baseiam na determinação de aflatoxinas B1, B2, G1 e G2 em amostras de amendoim, utilizando a técnica de espectroscopia de fluorescência molecular. O primeiro estudo tem o objetivo de avaliar a metodologia empregada para a quantificação de aflatoxinas totais em amendoins, utilizando o método clássico de validação fazendo-se o uso da calibração univariada. Os principais parâmetros de desempenho foram avaliados visando certificar a possibilidade de implementação desta metodologia em laboratórios. O segundo estudo está focado na separação e quantificação destas aflatoxinas com a aplicação combinada da espectrofluorimetria e de um método quimiométrico de segunda ordem (PARAFAC) utilizando a calibração multivariada. Esta técnica pode ser empregada como uma alternativa viável para a determinação de aflatoxinas B1, B2, G1 e G2 isoladamente, tradicionalmente é feito por cromatografia líquida de alta eficiência com detector de fluorescência. Porém, como estes analitos apresentam uma larga faixa de sobreposição espectral e as aflatoxinas (B1 e G1) possuem intensidade de sinal de fluorescência bem abaixo das demais, a separação e quantificação das quatro aflatoxinas foi inviável. O estudo foi retomado com a utilização das aflatoxinas B2 e G2 e os resultados alcançados foram satisfatórios. O método utilizado para a quantificação de aflatoxinas totais apresentou bons resultados, mostrando-se como uma importante ferramenta para a determinação destes analitos. Alem disso, comtempla perfeitamente o que é requerido pela legislação brasileira para a análise de aflatoxinas B1, B2, G1 e G2, que tem como exigência em laudos finais de análise a declaração do somatório, em g/kg, destas aflatoxinas, ou seja, sem a necessidade de quantifica-las separadamente / This research describes two case studies that are based on the determination of aflatoxins B1, B2, G1 and G2 in peanut samples, using the technique of molecular fluorescence spectroscopy. The first study aims to evaluate the methodology used for the quantification of total aflatoxins in peanuts, using the classical validation method by employing univariate calibration. The main performance parameters were evaluated to assure the ability to ensure implementation of this methodology in laboratories. The second study focuses on the separation and quantification of these aflatoxins with the combined application spectrofluorimetry and a chemometric method of second order (PARAFAC) using the multivariate calibration. This technique can be used as a viable alternative for the determination of aflatoxins B1, B2, G1 and G2 alone, which traditionally is done by high performance liquid chromatography with fluorescence detection. However, as these analytes exhibit a wide range of spectral overlap and the aflatoxins (B1 and G1) have intensity of fluorescence signal well below the others, the separation and quantification of the four aflatoxins was not feasible. The study was resumed with the use of aflatoxins B2 and G2 and the results were satisfactory.
The method used for quantification of total aflatoxins showed good results, showing itself as an important tool for the determination of these analytes. Furthermore, it contemplates exactly what is required by Brazilian legislation for the analysis of aflatoxins B1, B2, G1 and G2, which as a requirement in the final analysis reports is the statement of the sum, in g/kg, these aflatoxins, ie without the need to quantify them separately
|
6 |
Desenvolvimento de metodologia para a determinação de aflatoxinas em amostras de amendoim usando espectrofluorimetria e análise dos fatores paralelos (PARAFAC) / Development of methodology for the determination of aflatoxins in peanut samples using spectrofluorimetry and parallel factor analysisRobson Alves Luiz 21 March 2012 (has links)
Desenvolvimento de metodologia para a determinação de aflatoxinas em amostras de amendoim usando espectrofluorimetria e análise dos fatores paralelos (PARAFAC). 2012. 92 f. Dissertação (Mestrado em Engenharia Química) - Instituto de Química, Universidade do Estado do Rio de Janeiro , Rio de Janeiro, 2012. Neste trabalho de pesquisa são descritos dois estudos de caso que se baseiam na determinação de aflatoxinas B1, B2, G1 e G2 em amostras de amendoim, utilizando a técnica de espectroscopia de fluorescência molecular. O primeiro estudo tem o objetivo de avaliar a metodologia empregada para a quantificação de aflatoxinas totais em amendoins, utilizando o método clássico de validação fazendo-se o uso da calibração univariada. Os principais parâmetros de desempenho foram avaliados visando certificar a possibilidade de implementação desta metodologia em laboratórios. O segundo estudo está focado na separação e quantificação destas aflatoxinas com a aplicação combinada da espectrofluorimetria e de um método quimiométrico de segunda ordem (PARAFAC) utilizando a calibração multivariada. Esta técnica pode ser empregada como uma alternativa viável para a determinação de aflatoxinas B1, B2, G1 e G2 isoladamente, tradicionalmente é feito por cromatografia líquida de alta eficiência com detector de fluorescência. Porém, como estes analitos apresentam uma larga faixa de sobreposição espectral e as aflatoxinas (B1 e G1) possuem intensidade de sinal de fluorescência bem abaixo das demais, a separação e quantificação das quatro aflatoxinas foi inviável. O estudo foi retomado com a utilização das aflatoxinas B2 e G2 e os resultados alcançados foram satisfatórios. O método utilizado para a quantificação de aflatoxinas totais apresentou bons resultados, mostrando-se como uma importante ferramenta para a determinação destes analitos. Alem disso, comtempla perfeitamente o que é requerido pela legislação brasileira para a análise de aflatoxinas B1, B2, G1 e G2, que tem como exigência em laudos finais de análise a declaração do somatório, em g/kg, destas aflatoxinas, ou seja, sem a necessidade de quantifica-las separadamente / This research describes two case studies that are based on the determination of aflatoxins B1, B2, G1 and G2 in peanut samples, using the technique of molecular fluorescence spectroscopy. The first study aims to evaluate the methodology used for the quantification of total aflatoxins in peanuts, using the classical validation method by employing univariate calibration. The main performance parameters were evaluated to assure the ability to ensure implementation of this methodology in laboratories. The second study focuses on the separation and quantification of these aflatoxins with the combined application spectrofluorimetry and a chemometric method of second order (PARAFAC) using the multivariate calibration. This technique can be used as a viable alternative for the determination of aflatoxins B1, B2, G1 and G2 alone, which traditionally is done by high performance liquid chromatography with fluorescence detection. However, as these analytes exhibit a wide range of spectral overlap and the aflatoxins (B1 and G1) have intensity of fluorescence signal well below the others, the separation and quantification of the four aflatoxins was not feasible. The study was resumed with the use of aflatoxins B2 and G2 and the results were satisfactory.
The method used for quantification of total aflatoxins showed good results, showing itself as an important tool for the determination of these analytes. Furthermore, it contemplates exactly what is required by Brazilian legislation for the analysis of aflatoxins B1, B2, G1 and G2, which as a requirement in the final analysis reports is the statement of the sum, in g/kg, these aflatoxins, ie without the need to quantify them separately
|
7 |
Analyse de la matière organique et ses propriétés dans l’environnement naturel en spectroscopie de fluorescence 3D traitée par PARAFAC / Analysis of organic matter and its properties in natural environment on 3D-fluorescence treated by PARAFACZhao, Huiyu 25 February 2011 (has links)
Les matrices d’excitation et d’émission de fluorescence (MEEF) sont utilisées pour caractériser la matière organique naturelle (MON). Afin de mieux exploiter ces informations, un algorithme trilinéaire, PARAFAC, est employé. Après l’élimination des diffusions Rayleigh et Raman et la correction de l’effet d’écran, cette méthode permet de séparer les composants spectraux présents dans les MEEF.Ce travail présente deux études : la qualification et la quantification de la MON selon son origine environnementale et le calcul des constantes de complexation de la MON et du cuivre sous forme ionique.Les composants spectraux et leurs intensités relatives sont calculés par PARAFAC à partir 1146 échantillons regroupés suivant les missions, leur type de milieu, ou le niveau de salinité. Pour étudier ces composants, une nouvelle représentation spectrale est proposée afin de mettre en évidence leur variabilité spectrale. Les résultats montrent que le regroupement d’échantillons d’origine diverse conserve le recouvrement spectral global et les intensités relatives. Sur l’ensemble du domaine spectral, les zones correspondant aux substances humiques sont peu variables, comparées à la zone protéinique.La complexation des métaux par la MON est analysée par une technique combinant quatre outils : l’ajout logarithmique d’ions métalliques, la mesure de MEEF, la méthode PARAFAC et l’algorithme PROSECE. La mesure du quenching de fluorescence ne se limite pas seulement à la modélisation d’une intensité de fluorescence mais à celle de l’intensité relative de chaque composant PARAFAC surpassant ainsi les méthodes utilisées jusqu’à présent. Finalement, l’application de cette technique originale permet de quantifier les propriétés de complexation de la MON à l'aide d'un modèle de complexation utilisant 2 sites de complexation par composant en utilisant la totalité du signal de fluorescence. / Fluorescence excitation and emission matrices (EEM) are used to characterize natural organic matter (NOM). To make best use of this information, PARAFAC, a trilinear algorithm is employed. After removing Rayleigh and Raman scattering and correction of the inner filter effect, this method allows separating the spectral components present in MEEF.This work presents two studies: the characterization and quantification of NOM according to its origin and calculation of environmental complexation constants of NOM towards copper as ionic form.Spectral components and their relative intensities are calculated by PARAFAC from 1146 samples gathered according to the missions, the medium type, or the salinity level. To study these components, a new spectral representation is proposed in order to highlight their spectral variability. The results show that even when samples spectra of various origins are clustered, the overall spectral overlap and the relative intensities remain almost similar. On the whole spectral range, areas corresponding to humic substances are quite variable, compared to the protein zone.Metal complexation by NOM is analyzed by combination of four tools: metal ions logarithmic addition, MEEF measurement, the PARAFAC dissociation method and the PROSECE modelling algorithm. Fluorescence quenching measurement is not only limited to the modelling of fluorescence intensity but also to the relative intensity of each PARAFAC-dissociated component though surpassing the methods used so far. Finally, the application of this improving technique leads to quantify NOM complexation properties using a two-complexing sites complexation model for each PARAFAC-dissociated component by using the whole fluorescence signal.
|
8 |
KHATRI-RAO PRODUCTS AND CONDITIONS FOR THE UNIQUENESS OF PARAFAC SOLUTIONS FOR IxJxK ARRAYSBush, Heather Michele Clyburn 01 January 2006 (has links)
One of the differentiating features of PARAFAC decompositions is that, under certain conditions, unique solutions are possible. The search for uniqueness conditions for the PARAFAC Decomposition has a limited past, spanning only three decades. The complex structure of the problem and the need for tensor algebras or other similarly abstract characterizations provided a roadblock to the development of uniqueness conditions. Theoretically, the PARAFAC decomposition surpasses its bilinear counterparts in that it is possible to obtain solutions that do not suffer from the rotational problem. However, not all PARAFAC solutions will be constrained sufficiently so that the resulting decomposition is unique. The work of Kruskal, 1977, provides the most in depth investigation into the conditions for uniqueness, so much so that many have assumed, without formal proof, that his sufficient conditions were also necessary. Aided by the introduction of Khatri-Rao products to represent the PARAFAC decomposition, ten Berge and Sidiropoulos (2002) used the column spaces of Khatri-Rao products to provide the first evidence for countering the claim of necessity, identifying PARAFAC decompositions that were unique when Kruskals condition was not met. Moreover, ten Berge and Sidiropoulos conjectured that, with additional k-rank restrictions, a class of decompositions could be formed where Kruskals condition would be necessary and sufficient. Unfortunately, the column space argument of ten Berge and Sidiropoulos was limited in its application and failed to provide an explanation of why uniqueness occurred. On the other hand, the use of orthogonal complement spaces provided an alternative approach to evaluate uniqueness that would provide a much richer return than the use of column spaces for the investigation of uniqueness. The Orthogonal Complement Space Approach (OCSA), adopted here, would provide: (1) the answers to lingering questions about the occurrence of uniqueness, (2) evidence that necessity would require more than a restriction on k-rank, and (3) an approach that could be extended to cases beyond those investigated by ten Berge and Sidiropoulos.
|
9 |
Measuring The Influence Of Environmental Conditions On Dissolved Organic Matter Biodegradability And Optical Properties: A Combined Field And Laboratory StudyLandsman-Gerjoi, Maxwell 01 January 2019 (has links)
Dissolved organic matter (DOM) plays a vital role in biogeochemical processes and can flux CO2 to the atmosphere when labile fractions are degraded, hence DOM degradation is increasingly studied. Some studies have suggested that fluorescence-derived substrate characteristics are useful metrics for estimating bioavailability (as prerequisite condition for biodegradability), however, recent findings on soil organic matter emphasize the importance of ecosystem scale factors such as physical separation of substrate from soil microbial communities and soil physiochemical cycles driving organic matter stability. I extend this principle to soil derived DOM and hypothesize that such environmental conditions, covariant with season, land use and landscape position, impact the composition of soil DOM and activity and abundance of the microbial community, which together govern DOM biodegradability. As a result, DOM bioavailability may not reliably be predicted using substrate characteristics alone. To test these hypotheses, I assessed aqueous soil extracts for water extractable organic carbon (WEOC) content, biodegradability, microbial biomass and fluorescence spectroscopy on water extractable organic matter (WEOM) across a range of environmental conditions in northern Vermont, USA. My results indicate that changes in environmental conditions affect composition, quantity, and biodegradability of DOM. WEOC concentrations were highest in the fall and lowest in the summer, while no significant differences were found between land covers or landscape position, however, DOM biodegradability was significantly higher in the agricultural (AG) site across seasons. Despite a shift in utilized substrate from less aromatic DOM in summer to more aromatic DOM in winter, biodegradability was similar for all seasons. The only exception were cold temperature incubations where microbial activity was depressed, and processing was halted. These results indicate that bioavailability cannot be reliably predicted based on fluorescence-based metric alone, rather, my core findings illustrate a complex picture of how environmental conditions, landscape characteristics, and substrate composition interact to drive the biodegradability of labile carbon pools in the soil environment.
This thesis includes i) a background and comprehensive review of literature to inform the reader of any relevant topics, ii) a paper submitted for publication to Biogeochemistry (Chapter 2), and iii) supplemental information containing figures and tables pertinent to the paper.
|
10 |
Analyse de la matière organique et ses propriétés dans l'environnement naturel en spectroscopie de fluorescence 3D traitée par PARAFACZhao, Huiyu 25 February 2011 (has links) (PDF)
Les matrices d'excitation et d'émission de fluorescence (MEEF) sont utilisées pour caractériser la matière organique naturelle (MON). Afin de mieux exploiter ces informations, un algorithme trilinéaire, PARAFAC, est employé. Après l'élimination des diffusions Rayleigh et Raman et la correction de l'effet d'écran, cette méthode permet de séparer les composants spectraux présents dans les MEEF.Ce travail présente deux études : la qualification et la quantification de la MON selon son origine environnementale et le calcul des constantes de complexation de la MON et du cuivre sous forme ionique.Les composants spectraux et leurs intensités relatives sont calculés par PARAFAC à partir 1146 échantillons regroupés suivant les missions, leur type de milieu, ou le niveau de salinité. Pour étudier ces composants, une nouvelle représentation spectrale est proposée afin de mettre en évidence leur variabilité spectrale. Les résultats montrent que le regroupement d'échantillons d'origine diverse conserve le recouvrement spectral global et les intensités relatives. Sur l'ensemble du domaine spectral, les zones correspondant aux substances humiques sont peu variables, comparées à la zone protéinique.La complexation des métaux par la MON est analysée par une technique combinant quatre outils : l'ajout logarithmique d'ions métalliques, la mesure de MEEF, la méthode PARAFAC et l'algorithme PROSECE. La mesure du quenching de fluorescence ne se limite pas seulement à la modélisation d'une intensité de fluorescence mais à celle de l'intensité relative de chaque composant PARAFAC surpassant ainsi les méthodes utilisées jusqu'à présent. Finalement, l'application de cette technique originale permet de quantifier les propriétés de complexation de la MON à l'aide d'un modèle de complexation utilisant 2 sites de complexation par composant en utilisant la totalité du signal de fluorescence.
|
Page generated in 0.029 seconds