Hyperspectral subspace identification

Signal subspace identification is a crucial first step in many hyperspectral processing algorithms such as target detection, change detection, classification, and unmixing. The identification of this subspace enables a correct dimensionality reduction, yielding gains in algorithm performance and complexity and in data storage. This paper introduces a new minimum mean square error-based approach to infer the signal subspace in hyperspectral imagery. The method, which is termed hyperspectral signal identification by minimum error, is eigen decomposition based, unsupervised, and fully automatic (i.e., it does not depend on any tuning parameters). It first estimates the signal and noise correlation matrices and then selects the subset of eigenvalues that best represents the signal subspace in the least squared error sense. State-of-the-art performance of the proposed method is illustrated by using simulated and real hyperspectral images.

Signal Subspace Identification in Hyperspectral Linear Mixtures

Chpater in Book Proceedings with Peer Review Second Iberian Conference, IbPRIA 2005, Estoril, Portugal, June 7-9, 2005, Proceedings, Part II

Reading performance in children with visual function anomalies

Aims - To compare reading performance in children with and without visual function anomalies and identify the influence of abnormal visual function and other variables in reading ability. Methods - A cross-sectional study was carried in 110 children of school age (6-11 years) with Abnormal Visual Function (AVF) and 562 children with Normal Visual Function (NVF). An orthoptic assessment (visual acuity, ocular alignment, near point of convergence and accommodation, stereopsis and vergences) and autorefraction was carried out. Oral reading was analyzed (list of 34 words). Number of errors, accuracy (percentage of success) and reading speed (words per minute - wpm) were used as reading indicators. Sociodemographic information from parents (n=670) and teachers (n=34) was obtained. Results - Children with AVF had a higher number of errors (AVF=3.00 errors; NVF=1.00 errors; p<0.001), a lower accuracy (AVF=91.18%; NVF=97.06%; p<0.001) and reading speed (AVF=24.71 wpm; NVF=27.39 wpm; p=0.007). Reading speed in the 3rd school grade was not statistically different between the two groups (AVF=31.41 wpm; NVF=32.54 wpm; p=0.113). Children with uncorrected hyperopia (p=0.003) and astigmatism (p=0.019) had worst reading performance. Children in 2nd, 3rd, or 4th grades presented a lower risk of having reading impairment when compared with the 1st grade. Conclusion - Children with AVF had reading impairment in the first school grade. It seems that reading abilities have a wide variation and this disparity lessens in older children. The slow reading characteristics of the children with AVF are similar to dyslexic children, which suggest the need for an eye evaluation before classifying the children as dyslexic.

Molecular epidemiology of Aspergillus collected from cystic fibrosis patients

Background - Aspergillus respiratory infection is a common complication in cystic fibrosis (CF) and is associated with loss of pulmonary function and allergic disease. Methods - Fifty-three Aspergillus isolates recovered from CF patients were identified to species by Internal Transcribed Spacer Region (ITS), β-tubulin, and calmodulin sequencing. Results - Three species complexes (Terrei, Nigri, and Fumigati) were found. Identification to species level gave a single Aspergillus terreus sensu stricto, one Aspergillus niger sensu stricto and 51 Aspergillus fumigatus sensu stricto isolates. No cryptic species were found. Conclusions - To our knowledge, this is the first prospective study of Aspergillus species in CF using molecular methods. The paucity of non-A. fumigatus and of cryptic species of A. fumigatus suggests a special association of A. fumigatus sensu stricto with CF airways, indicating it likely displays unique characteristics making it suitable for chronic residence in that milieu. These findings could refine an epidemiologic and therapeutic approach geared to this pathogen.

Modelo de competências dos enfermeiros com funções de gestão em Portugal: estudo exploratório

Nas últimas décadas assistimos a transformações económicas, tecnológicas, políticas e sociais, que influenciaram diretamente o modo de pensar e agir nas organizações. O conceito de competências, com uma valorização crescente, surge como uma alternativa à abordagem da gestão de recursos humanos por funções, respondendo aos desafios atuais do mercado: necessidade de flexibilidade, de adaptação a mudanças contínuas, exigências crescentes do mercado e competitividade das organizações nesse mercado. A área da saúde, e concretamente a profissão de Enfermagem também tem evoluído, surgindo em 2009, uma nova forma de operacionalizar a carreira destes profissionais. No que diz respeito aos enfermeiros com funções de gestão, o conteúdo funcional está descrito, contudo, não existe uma definição clara das competências requeridas para estes profissionais. Este trabalho de investigação, de cariz exploratório, utilizando uma metodologia qualitativa, pretendeu propor uma estratégia de definição de um modelo de competências para os enfermeiros com funções de gestão em Portugal. Para isso, definimos categorias de competências, através da análise da literatura e da legislação. Seguiu-se a realização de entrevistas a um painel de doze peritos, e uma análise de conteúdo dos dados (categorização do tipo misto). Procedemos a uma comparação da recolha empírica de competências com as da recolha teórica, e definimos uma lista de 10 competências para as funções de gestão dos enfermeiros: Competências Técnicas de Gestão; Competências Interpessoais; Comunicação; Gestão de Recursos Humanos; Pensamento Crítico; Conhecimento de Políticas de Saúde; Competências Técnicas de Enfermagem; Organização e Planeamento; Trabalho de Equipa; Preocupação pela Qualidade. De forma a complementar o estudo, pretendemos identificar a perceção das lacunas de competências nos enfermeiros com funções de gestão, e identificar os processos de desenvolvimento de competências considerados mais relevantes para estes profissionais. As lacunas identificadas nas competências dos atuais enfermeiros com funções de gestão, face às mais valorizadas, são reduzidas e dispersas, pelo que consideramos pouco significativas. A forma de desenvolvimento de competências mais valorizado pelo painel de peritos foi a formação (académica e em contexto profissional). Foi também realçada a importância do empenho individual neste processo, assim como a avaliação de competências antes dos enfermeiros assumirem funções de gestão.Consideramos que esta investigação traz contributos quer para a literatura da Gestão por Competências, quer para a literatura da definição de competências das funções dos enfermeiros com funções de gestão, quer para a profissão de enfermagem, (nomeadamente, para as funções de gestão dos enfermeiros), quer para o próprio SNS, já que faz algumas propostas e sugestões para a evolução das práticas de gestão de pessoas.

Contribution of hospital effluents to the load of pharmaceuticals in urban wastewaters: Identification of ecologically relevant pharmaceuticals

The impact of effluent wastewaters from four different hospitals: a university (1456 beds), a general (350 beds), a pediatric (110 beds) and a maternity hospital (96 beds), which are conveyed to the same wastewater treatment plant (WWTP), was evaluated in the receiving urban wastewaters. The occurrence of 78 pharmaceuticals belonging to several therapeutic classes was assessed in hospital effluents and WWTP wastewaters (influent and effluent) as well as the contribution of each hospital in WWTP influent in terms of pharmaceutical load. Results indicate that pharmaceuticals are widespread pollutants in both hospital and urban wastewaters. The contribution of hospitals to the input of pharmaceuticals in urban wastewaters widely varies, according to their dimension. The estimated total mass loadings were 306 g d− 1 for the university hospital, 155 g d− 1 for the general one, 14 g d− 1 for the pediatric hospital and 1.5 g d− 1 for the maternity hospital, showing that the biggest hospitals have a greater contribution to the total mass load of pharmaceuticals. Furthermore, analysis of individual contributions of each therapeutic group showed that NSAIDs, analgesics and antibiotics are among the groups with the highest inputs. Removal efficiency can go from over 90% for pharmaceuticals like acetaminophen and ibuprofen to not removal for β-blockers and salbutamol. Total mass load of pharmaceuticals into receiving surface waters was estimated between 5 and 14 g/d/1000 inhabitants. Finally, the environmental risk posed by pharmaceuticals detected in hospital and WWTP effluents was assessed by means of hazard quotients toward different trophic levels (algae, daphnids and fish). Several pharmaceuticals present in the different matrices were identified as potentially hazardous to aquatic organisms, showing that especial attention should be paid to antibiotics such as ciprofloxacin, ofloxacin, sulfamethoxazole, azithromycin and clarithromycin, since their hazard quotients in WWTP effluent revealed that they could pose an ecotoxicological risk to algae.

Brewer's spent grain from different types of malt: Evaluation of the antioxidant activity and identification of the major phenolic compounds

The antioxidant activity and phenolic composition of brewer's spent grain (BSG) extracts obtained by microwave-assisted extraction from twomalt types (light and darkmalts) were investigated. The total phenolic content (TPC) and antioxidant activity among the light BSG extracts (pilsen, melano, melano 80 and carared)were significantly different (p b 0.05) compared to dark extracts (chocolate and black types), with the pilsen BSG showing higher TPC (20 ± 1 mgGAE/g dry BSG). In addition, the antioxidant activity assessed by 2,2-diphenyl- 1-picrylhydrazyl, 2,2′-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) and deoxyribose assays decreased as a result of increasing kilning temperatures in the following order: pilsen N melano N melano 80 N carared N chocolate N black. HPLC-DAD/ESI-MS/MS analysis indicated the presence of phenolic acids, such as ferulic, p-coumaric and syringic acids, as well as several isomeric ferulate dehydrodimers and one dehydrotrimer. Chocolate and black extracts, obtained frommalts submitted to the highest kilning temperatures, showed the lowest levels of ferulic and p-coumaric acids. These results suggested that BSG extracts from pilsen malt might be used as an inexpensive and good natural source of antioxidants with potential interest for the food, pharmaceutical and/or cosmetic industries after purification.

PAH air pollution at a Portuguese urban area: carcinogenic risks and sources identification

This study aimed to characterize air pollution and the associated carcinogenic risks of polycyclic aromatic hydrocarbon (PAHs) at an urban site, to identify possible emission sources of PAHs using several statistical methodologies, and to analyze the influence of other air pollutants and meteorological variables on PAH concentrations.The air quality and meteorological data were collected in Oporto, the second largest city of Portugal. Eighteen PAHs (the 16 PAHs considered by United States Environment Protection Agency (USEPA) as priority pollutants, dibenzo[a,l]pyrene, and benzo[j]fluoranthene) were collected daily for 24 h in air (gas phase and in particles) during 40 consecutive days in November and December 2008 by constant low-flow samplers and using polytetrafluoroethylene (PTFE) membrane filters for particulate (PM10 and PM2.5 bound) PAHs and pre-cleaned polyurethane foam plugs for gaseous compounds. The other monitored air pollutants were SO2, PM10, NO2, CO, and O3; the meteorological variables were temperature, relative humidity, wind speed, total precipitation, and solar radiation. Benzo[a]pyrene reached a mean concentration of 2.02 ngm−3, surpassing the EU annual limit value. The target carcinogenic risks were equal than the health-based guideline level set by USEPA (10−6) at the studied site, with the cancer risks of eight PAHs reaching senior levels of 9.98×10−7 in PM10 and 1.06×10−6 in air. The applied statistical methods, correlation matrix, cluster analysis, and principal component analysis, were in agreement in the grouping of the PAHs. The groups were formed according to their chemical structure (number of rings), phase distribution, and emission sources. PAH diagnostic ratios were also calculated to evaluate the main emission sources. Diesel vehicular emissions were the major source of PAHs at the studied site. Besides that source, emissions from residential heating and oil refinery were identified to contribute to PAH levels at the respective area. Additionally, principal component regression indicated that SO2, NO2, PM10, CO, and solar radiation had positive correlation with PAHs concentrations, while O3, temperature, relative humidity, and wind speed were negatively correlated.

Comparação de dois protocolos de reabilitação após reconstrução cirúrgica do ligamento cruzado anterior

Mestrado em Fisioterapia

Experimental design methodologies for the identification of Michaelis-Menten type kinetics

The main objective of this work was to investigate the application of experimental design techniques for the identification of Michaelis-Menten kinetic parameters. More specifically, this study attempts to elucidate the relative advantages/disadvantages of employing complex experimental design techniques in relation to equidistant sampling when applied to different reactor operation modes. All studies were supported by simulation data of a generic enzymatic process that obeys to the Michaelis-Menten kinetic equation. Different aspects were investigated, such as the influence of the reactor operation mode (batch, fed-batch with pulse wise feeding and fed-batch with continuous feeding) and the experimental design optimality criteria on the effectiveness of kinetic parameters identification. The following experimental design optimality criteria were investigated: 1) minimization of the sum of the diagonal of the Fisher information matrix (FIM) inverse (A-criterion), 2) maximization of the determinant of the FIM (D-criterion), 3) maximization of the smallest eigenvalue of the FIM (E-criterion) and 4) minimization of the quotient between the largest and the smallest eigenvalue (modified E-criterion). The comparison and assessment of the different methodologies was made on the basis of the Cramér-Rao lower bounds (CRLB) error in respect to the parameters vmax and Km of the Michaelis-Menten kinetic equation. In what concerns the reactor operation mode, it was concluded that fed-batch (pulses) is better than batch operation for parameter identification. When the former operation mode is adopted, the vmax CRLB error is lowered by 18.6 % while the Km CRLB error is lowered by 26.4 % when compared to the batch operation mode. Regarding the optimality criteria, the best method was the A-criterion, with an average vmax CRLB of 6.34 % and 5.27 %, for batch and fed-batch (pulses), respectively, while presenting a Km’s CRLB of 25.1 % and 18.1 %, for batch and fed-batch (pulses), respectively. As a general conclusion of the present study, it can be stated that experimental design is justified if the starting parameters CRLB errors are inferior to 19.5 % (vmax) and 45% (Km), for batch processes, and inferior to 42 % and to 50% for fed-batch (pulses) process. Otherwise equidistant sampling is a more rational decision. This conclusion clearly supports that, for fed-batch operation, the use of experimental design is likely to largely improve the identification of Michaelis-Menten kinetic parameters.

Demo: automatic personal identification system for security in critical services: a case study

The demonstration proposal moves from the capabilities of a wireless biometric badge [4], which integrates a localization and tracking service along with an automatic personal identification mechanism, to show how a full system architecture is devised to enable the control of physical accesses to restricted areas. The system leverages on the availability of a novel IEEE 802.15.4/Zigbee Cluster Tree network model, on enhanced security levels and on the respect of all the users' privacy issues.

Memória de longo prazo nos retornos acionistas dos índices de referência da euronext, implicações para a hipótese de mercados eficientes e contributo fractal para aperfeiçoamento do capital asset pricing model

Não existe uma definição única de processo de memória de longo prazo. Esse processo é geralmente definido como uma série que possui um correlograma decaindo lentamente ou um espectro infinito de frequência zero. Também se refere que uma série com tal propriedade é caracterizada pela dependência a longo prazo e por não periódicos ciclos longos, ou que essa característica descreve a estrutura de correlação de uma série de longos desfasamentos ou que é convencionalmente expressa em termos do declínio da lei-potência da função auto-covariância. O interesse crescente da investigação internacional no aprofundamento do tema é justificado pela procura de um melhor entendimento da natureza dinâmica das séries temporais dos preços dos ativos financeiros. Em primeiro lugar, a falta de consistência entre os resultados reclama novos estudos e a utilização de várias metodologias complementares. Em segundo lugar, a confirmação de processos de memória longa tem implicações relevantes ao nível da (1) modelação teórica e econométrica (i.e., dos modelos martingale de preços e das regras técnicas de negociação), (2) dos testes estatísticos aos modelos de equilíbrio e avaliação, (3) das decisões ótimas de consumo / poupança e de portefólio e (4) da medição de eficiência e racionalidade. Em terceiro lugar, ainda permanecem questões científicas empíricas sobre a identificação do modelo geral teórico de mercado mais adequado para modelar a difusão das séries. Em quarto lugar, aos reguladores e gestores de risco importa saber se existem mercados persistentes e, por isso, ineficientes, que, portanto, possam produzir retornos anormais. O objetivo do trabalho de investigação da dissertação é duplo. Por um lado, pretende proporcionar conhecimento adicional para o debate da memória de longo prazo, debruçando-se sobre o comportamento das séries diárias de retornos dos principais índices acionistas da EURONEXT. Por outro lado, pretende contribuir para o aperfeiçoamento do capital asset pricing model CAPM, considerando uma medida de risco alternativa capaz de ultrapassar os constrangimentos da hipótese de mercado eficiente EMH na presença de séries financeiras com processos sem incrementos independentes e identicamente distribuídos (i.i.d.). O estudo empírico indica a possibilidade de utilização alternativa das obrigações do tesouro (OT’s) com maturidade de longo prazo no cálculo dos retornos do mercado, dado que o seu comportamento nos mercados de dívida soberana reflete a confiança dos investidores nas condições financeiras dos Estados e mede a forma como avaliam as respetiva economias com base no desempenho da generalidade dos seus ativos. Embora o modelo de difusão de preços definido pelo movimento Browniano geométrico gBm alegue proporcionar um bom ajustamento das séries temporais financeiras, os seus pressupostos de normalidade, estacionariedade e independência das inovações residuais são adulterados pelos dados empíricos analisados. Por isso, na procura de evidências sobre a propriedade de memória longa nos mercados recorre-se à rescaled-range analysis R/S e à detrended fluctuation analysis DFA, sob abordagem do movimento Browniano fracionário fBm, para estimar o expoente Hurst H em relação às séries de dados completas e para calcular o expoente Hurst “local” H t em janelas móveis. Complementarmente, são realizados testes estatísticos de hipóteses através do rescaled-range tests R/S , do modified rescaled-range test M - R/S e do fractional differencing test GPH. Em termos de uma conclusão única a partir de todos os métodos sobre a natureza da dependência para o mercado acionista em geral, os resultados empíricos são inconclusivos. Isso quer dizer que o grau de memória de longo prazo e, assim, qualquer classificação, depende de cada mercado particular. No entanto, os resultados gerais maioritariamente positivos suportam a presença de memória longa, sob a forma de persistência, nos retornos acionistas da Bélgica, Holanda e Portugal. Isto sugere que estes mercados estão mais sujeitos a maior previsibilidade (“efeito José”), mas também a tendências que podem ser inesperadamente interrompidas por descontinuidades (“efeito Noé”), e, por isso, tendem a ser mais arriscados para negociar. Apesar da evidência de dinâmica fractal ter suporte estatístico fraco, em sintonia com a maior parte dos estudos internacionais, refuta a hipótese de passeio aleatório com incrementos i.i.d., que é a base da EMH na sua forma fraca. Atendendo a isso, propõem-se contributos para aperfeiçoamento do CAPM, através da proposta de uma nova fractal capital market line FCML e de uma nova fractal security market line FSML. A nova proposta sugere que o elemento de risco (para o mercado e para um ativo) seja dado pelo expoente H de Hurst para desfasamentos de longo prazo dos retornos acionistas. O expoente H mede o grau de memória de longo prazo nos índices acionistas, quer quando as séries de retornos seguem um processo i.i.d. não correlacionado, descrito pelo gBm(em que H = 0,5 , confirmando- se a EMH e adequando-se o CAPM), quer quando seguem um processo com dependência estatística, descrito pelo fBm(em que H é diferente de 0,5, rejeitando-se a EMH e desadequando-se o CAPM). A vantagem da FCML e da FSML é que a medida de memória de longo prazo, definida por H, é a referência adequada para traduzir o risco em modelos que possam ser aplicados a séries de dados que sigam processos i.i.d. e processos com dependência não linear. Então, estas formulações contemplam a EMH como um caso particular possível.

Automatic learning for the classification of chemical reactions and in statistical thermodynamics

This Thesis describes the application of automatic learning methods for a) the classification of organic and metabolic reactions, and b) the mapping of Potential Energy Surfaces(PES). The classification of reactions was approached with two distinct methodologies: a representation of chemical reactions based on NMR data, and a representation of chemical reactions from the reaction equation based on the physico-chemical and topological features of chemical bonds. NMR-based classification of photochemical and enzymatic reactions. Photochemical and metabolic reactions were classified by Kohonen Self-Organizing Maps (Kohonen SOMs) and Random Forests (RFs) taking as input the difference between the 1H NMR spectra of the products and the reactants. The development of such a representation can be applied in automatic analysis of changes in the 1H NMR spectrum of a mixture and their interpretation in terms of the chemical reactions taking place. Examples of possible applications are the monitoring of reaction processes, evaluation of the stability of chemicals, or even the interpretation of metabonomic data. A Kohonen SOM trained with a data set of metabolic reactions catalysed by transferases was able to correctly classify 75% of an independent test set in terms of the EC number subclass. Random Forests improved the correct predictions to 79%. With photochemical reactions classified into 7 groups, an independent test set was classified with 86-93% accuracy. The data set of photochemical reactions was also used to simulate mixtures with two reactions occurring simultaneously. Kohonen SOMs and Feed-Forward Neural Networks (FFNNs) were trained to classify the reactions occurring in a mixture based on the 1H NMR spectra of the products and reactants. Kohonen SOMs allowed the correct assignment of 53-63% of the mixtures (in a test set). Counter-Propagation Neural Networks (CPNNs) gave origin to similar results. The use of supervised learning techniques allowed an improvement in the results. They were improved to 77% of correct assignments when an ensemble of ten FFNNs were used and to 80% when Random Forests were used. This study was performed with NMR data simulated from the molecular structure by the SPINUS program. In the design of one test set, simulated data was combined with experimental data. The results support the proposal of linking databases of chemical reactions to experimental or simulated NMR data for automatic classification of reactions and mixtures of reactions. Genome-scale classification of enzymatic reactions from their reaction equation. The MOLMAP descriptor relies on a Kohonen SOM that defines types of bonds on the basis of their physico-chemical and topological properties. The MOLMAP descriptor of a molecule represents the types of bonds available in that molecule. The MOLMAP descriptor of a reaction is defined as the difference between the MOLMAPs of the products and the reactants, and numerically encodes the pattern of bonds that are broken, changed, and made during a chemical reaction. The automatic perception of chemical similarities between metabolic reactions is required for a variety of applications ranging from the computer validation of classification systems, genome-scale reconstruction (or comparison) of metabolic pathways, to the classification of enzymatic mechanisms. Catalytic functions of proteins are generally described by the EC numbers that are simultaneously employed as identifiers of reactions, enzymes, and enzyme genes, thus linking metabolic and genomic information. Different methods should be available to automatically compare metabolic reactions and for the automatic assignment of EC numbers to reactions still not officially classified. In this study, the genome-scale data set of enzymatic reactions available in the KEGG database was encoded by the MOLMAP descriptors, and was submitted to Kohonen SOMs to compare the resulting map with the official EC number classification, to explore the possibility of predicting EC numbers from the reaction equation, and to assess the internal consistency of the EC classification at the class level. A general agreement with the EC classification was observed, i.e. a relationship between the similarity of MOLMAPs and the similarity of EC numbers. At the same time, MOLMAPs were able to discriminate between EC sub-subclasses. EC numbers could be assigned at the class, subclass, and sub-subclass levels with accuracies up to 92%, 80%, and 70% for independent test sets. The correspondence between chemical similarity of metabolic reactions and their MOLMAP descriptors was applied to the identification of a number of reactions mapped into the same neuron but belonging to different EC classes, which demonstrated the ability of the MOLMAP/SOM approach to verify the internal consistency of classifications in databases of metabolic reactions. RFs were also used to assign the four levels of the EC hierarchy from the reaction equation. EC numbers were correctly assigned in 95%, 90%, 85% and 86% of the cases (for independent test sets) at the class, subclass, sub-subclass and full EC number level,respectively. Experiments for the classification of reactions from the main reactants and products were performed with RFs - EC numbers were assigned at the class, subclass and sub-subclass level with accuracies of 78%, 74% and 63%, respectively. In the course of the experiments with metabolic reactions we suggested that the MOLMAP / SOM concept could be extended to the representation of other levels of metabolic information such as metabolic pathways. Following the MOLMAP idea, the pattern of neurons activated by the reactions of a metabolic pathway is a representation of the reactions involved in that pathway - a descriptor of the metabolic pathway. This reasoning enabled the comparison of different pathways, the automatic classification of pathways, and a classification of organisms based on their biochemical machinery. The three levels of classification (from bonds to metabolic pathways) allowed to map and perceive chemical similarities between metabolic pathways even for pathways of different types of metabolism and pathways that do not share similarities in terms of EC numbers. Mapping of PES by neural networks (NNs). In a first series of experiments, ensembles of Feed-Forward NNs (EnsFFNNs) and Associative Neural Networks (ASNNs) were trained to reproduce PES represented by the Lennard-Jones (LJ) analytical potential function. The accuracy of the method was assessed by comparing the results of molecular dynamics simulations (thermal, structural, and dynamic properties) obtained from the NNs-PES and from the LJ function. The results indicated that for LJ-type potentials, NNs can be trained to generate accurate PES to be used in molecular simulations. EnsFFNNs and ASNNs gave better results than single FFNNs. A remarkable ability of the NNs models to interpolate between distant curves and accurately reproduce potentials to be used in molecular simulations is shown. The purpose of the first study was to systematically analyse the accuracy of different NNs. Our main motivation, however, is reflected in the next study: the mapping of multidimensional PES by NNs to simulate, by Molecular Dynamics or Monte Carlo, the adsorption and self-assembly of solvated organic molecules on noble-metal electrodes. Indeed, for such complex and heterogeneous systems the development of suitable analytical functions that fit quantum mechanical interaction energies is a non-trivial or even impossible task. The data consisted of energy values, from Density Functional Theory (DFT) calculations, at different distances, for several molecular orientations and three electrode adsorption sites. The results indicate that NNs require a data set large enough to cover well the diversity of possible interaction sites, distances, and orientations. NNs trained with such data sets can perform equally well or even better than analytical functions. Therefore, they can be used in molecular simulations, particularly for the ethanol/Au (111) interface which is the case studied in the present Thesis. Once properly trained, the networks are able to produce, as output, any required number of energy points for accurate interpolations.

Characterization and identification of printed objects

A study about the physical appearance of pre-photographic, photomechanical, photographic and digital positive reflective prints was made, relating the obtained images with the history, materials and technology used to create them. The studied samples are from the Image Permanence Institute (IPI) study collection. The digital images were obtained using a digital SLR on a copystand and a compound light microscope, with different lighting angles (0º, 45ºand 90º) and magnifications from overall views on the copystand down to a 20x objective lens on the microscope. Most of these images were originally created by IPI for www.digitalsamplebook.org, a web tool for teaching print identification, and will be used on the www.graphicsatlas.org website, along with textual information on identification, technology and history information about these reproduction processes.

Penalty fuzzy function for derivative-free optimization

Penalty and Barrier methods are normally used to solve Nonlinear Optimization Problems constrained problems. The problems appear in areas such as engineering and are often characterised by the fact that involved functions (objective and constraints) are non-smooth and/or their derivatives are not know. This means that optimization methods based on derivatives cannot net used. A Java based API was implemented, including only derivative-free optimizationmethods, to solve both constrained and unconstrained problems, which includes Penalty and Barriers methods. In this work a new penalty function, based on Fuzzy Logic, is presented. This function imposes a progressive penalization to solutions that violate the constraints. This means that the function imposes a low penalization when the violation of the constraints is low and a heavy penalisation when the violation is high. The value of the penalization is not known in beforehand, it is the outcome of a fuzzy inference engine. Numerical results comparing the proposed function with two of the classic penalty/barrier functions are presented. Regarding the presented results one can conclude that the prosed penalty function besides being very robust also exhibits a very good performance.