Thesis (M.S.)--Worcester Polytechnic Institute. / Keywords: Diabetes; imaging; phosphorescence; retina. Includes bibliographical references (p. 69-70).

Sous réserve d’une énergie cinétique suffisamment faible, un neutron libre peut être piégé matériellement ou magnétiquement de sorte à garantir son confinement au sein d’un volume défini. Cette caractéristique permet l’étude de plusieurs paramètres, notamment de son temps de vie moyen. L’expérience HOPE, piège magnétique de neutrons ultra-froids mis en œuvre à l’Institut Laue Langevin à Grenoble, vise à fournir une valeur précise de ce temps de vie au travers de différentes méthodes. L’une d’entre elles consiste à observer les électrons émis par la décroissance bêta du neutron. Le détecteur SECOND a été spécifiquement conçu pour permettre le comptage de ces électrons au sein de l’expérience HOPE. La grande difficulté de ce projet réside dans le faible taux de comptage des électrons attendu, qui nécessite la discrimination des rayonnements parasites. Dans ce but, SECOND est constitué de deux étages de détection, dont le principal, un phoswich de scintillateurs plastiques, a donné des résultats probants lors de premiers tests fonctionnels à basse température ; la différenciation des événements induits par des muons cosmiques est efficace dans 98 % des cas, et tout porte à croire qu’elle sera considérablement améliorée par l’utilisation d’un système d’acquisition adapté à l’application souhaitée. / Considering a low enough kinetic energy, a free neutron can be materially or magnetically trapped in a defined volume. This trapping allows experimenters to study the neutron and its characteristics, and in this case, to measure its mean lifetime. The HOPE experiment commissioned at Laue Langevin Institute in Grenoble is aimed at providing a 1 %- accuracy value. One way to measure lifetime is to record every single neutron beta decay occurring in the trap by counting the emitted electrons. The detector SECOND has been specifically designed to fulfill this goal within HOPE but also to discriminate other types of particles that induce false events. The latter argument is the reason for the two detection stages SECOND is composed of. The plastic scintillators phoswich constitutes the main part of the detector and has been successfully operated during preliminary tests at low temperature. The rejection rate of cosmic muons events is about 98 %, and this value can be drastically enhanced using a more suitable data acquisition system.

Title: Measurement of the B0 d meson lifetime at ATLAS detector Author: Pavel Řezníček Department: Institute of Particle and Nuclear Physics Supervisor of the doctoral thesis: Dr. Zdeněk Doležal Supervisor's e-mail address: Zdenek.Dolezal@mff.cuni.cz Abstract: The lifetime of B0 d mesons is determined from their decays B0 d → J/ψK∗0 reconstructed in ATLAS experiment at the LHC using pp collision data at a center-of-mass energy of 7 TeV and corresponding to integrated luminosity of 40 pb−1 . The lifetime, extracted from the simultaneous unbinned maximum likelihood mass-lifetime fit, is 1.51±0.04 (stat.)±0.04 (syst.) ps. A total number of 2750±90 (stat.) signal B0 d decays are observed in the measurement, with a fit- ted B0 d mass of 5363.7 ± 1.2 (stat.) MeV. Both the extracted B0 d meson mass and lifetime are within the determined errors consistent with the world average values. Although the achieved precision is still significantly lower than the one of the world average value, the measurement successfully tested the feasibility of the fit-method and allowed to cross-check ATLAS detector performance. Keywords: CERN, LHC, ATLAS, B-physics, B-hadron, Lifetime

Amyloid $\beta$ (1-42) (A$\beta$42) is a seminal neuropathic agent in Alzheimer’s disease (AD), a multifaceted neurodegenerative disorder for which no preventative measures or disease modifying therapies currently exist. Aggregation of this peptide plays a key role in the synaptic dysfunction and neuronal death associated with the disease. Perturbing the aggregation process, therefore, represents a key strategy for the development of new AD therapeutics. A variety of issues with current screening methods, including lack of reproducibility, high reagent consumption and spectral interference from the test molecules, can limit efforts to identify new small molecule inhibitors. Furthermore, the lack of robust, time- and cost-efficient methods for screening compounds in cellular or in vivo models limits the throughput with which seemingly active small molecules can be validated and prioritised. Herein, this thesis describes efforts to overcome such limitations through the development of a unified in vitro to in vivo assay system, in which hits identified in the ‘nanoFLIM’ microfluidic-based assay can quickly be tested in cellular and whole organism disease models. The assay platform designed relies on the use of an amyloid aggregation fluorescence lifetime sensor. A$\beta$42 aggregation is monitored by changes in the fluorescence lifetime of an attached fluorophore, which is significantly quenched upon amyloid formation. To take advantage of the benefits associated with miniaturisation, an in vitro microfluidic platform was employed. A microfluidic chip capable of trapping 110 precisely ordered droplets was designed, allowing for increased sample size and greatly lowering reagent consumption relative to conventional assay formats. Optimisation of the lifetime sensor technique permitted real-time compound screening in SH-SY5Y neuroblastoma cells, as well as in disease model Caenorhabditis elegans (C. elegans). To demonstrate the potential of this assay, a selection of novel chemical libraries developed in the Spring research group was screened, resulting in the identification of a key library of interest. The inhibitory activity of the lead compound from this collection was validated using a variety of biophysical tests, and was also shown to suppress amyloid aggregation in the live cell fluorescence lifetime sensor assay, as well as in whole organism disease model C. elegans. Whilst assay development was underway, additional screening of structurally diverse chemical libraries was performed using a conventional Thioflavin T spectroscopic assay. Such work identified another molecular scaffold capable of exerting a strong inhibitory effect against A$\beta$42 aggregation. A selection of analogues was synthesised to improve the in vivo profile of this library, giving rise to a second lead inhibitory compound. The activity of this compound was subsequently validated in biophysical and cellular tests, and was also tested in disease model Drosophila melanogaster. The aggregation of A$\beta$42 lies at the root of Alzheimer’s disease. In light of the relatively few drug candidates in clinical trials for this disorder, the development of improved translational screening approaches and continued screening of novel chemical libraries is necessary to identify new potential therapeutics. In this study, an in vitro to in vivo fluorescence lifetime imaging assay has been established. Using this assay system and conventional screening approaches, two A$\beta$42 aggregation inhibitors have been identified and validated. These represent promising candidates for the development of new AD therapeutic agents, or for use as molecular probes to further dissect the mechanisms underlying this devastating disease.

Recentes pesquisas no campo do Marketing têm apresentado resultados demonstrando que estratégias baseadas em elevar o Valor Vitalício dos Clientes (CLV) podem ter um impacto positivo no valor ao acionista. Em consequência disso, o Customer Equity, equivalente à soma dos CLVs dos clientes, também chamado de valor da base de clientes de uma empresa, vem se estabelecendo como métrica capaz de comprovar o impacto do Marketing no valor das empresas. A lógica que sustenta esta relação está no fato de que o valor das ações está baseado na expectativa de fluxo de caixa futuro, que provêm da base de clientes. O cálculo do CLV dos clientes também apresenta uma importante utilidade gerencial como balizador da correta alocação de recursos de Marketing, indicando os clientes mais lucrativos no longo prazo. A partir do modelo de Kumar e Shah (2009), esta dissertação realiza o cálculo do Customer Equity de uma empresa do setor financeiro a partir de informações individuais das transações dos clientes. O CLV é calculado individualmente, permitindo identificar o perfil dos clientes mais rentáveis no longo-prazo. Também é testada a relação entre Customer Equity e capitalização de mercado ao longo de 31 meses, sem comprovação de associação significativa entre as variáveis, mas havendo um sinal de relação positiva entre elas. Os resultados são discutidos, e, por fim, são analisadas implicações gerenciais e sugestões de pesquisas futuras. / Recent research in the field of Marketing have presented results demonstrating that strategies based on raising the Customer Lifetime Value (CLV) can have a positive impact on shareholder value. As a result, the Customer Equity, the sum of the customer’s CLVs, also called a customer’s base value of a company, has established itself as a metric able to demonstrate the impact of Marketing on the value of companies. The logic that underlies this relationship is in the fact that the value of the shares is based on expected future cash flows that stem from the customer base. Calculating customer’s CLV also has an important management utility of indicate the proper allocation of Marketing resources, identifying the most profitable customers in the long run. From the model of Kumar and Shah (2009), this paper performs the calculation of the Customer Equity of a financial institution from individual details of customer’s transactions. The CLV is calculated individually, allowing to identify the profile of the most profitable customers in the long run. Is also tested the relationship between Customer Equity and market capitalization over 31 months, without being proven a significant association between variables, but having a sign of a positive relationship between them. The results are discussed, and finally, managerial implications and suggestions for future research are analyzed.

Printed electronics is a combination of electronics and printing technologies commonly used in the publication industry such as screen, inkjet, and roll to roll printing. The measurements of conductive film particularly the conductive paste is the main objective of this thesis. The conductive paste consists of conductive filler, adhesive and solvent. Each component affects the electrical, and mechanical properties of the finished conductive film product. The measurements of conductive film have three field of study. The first category is the lifetime performance measurement of conductive film using environmental testing. A screen printed carbon, silver and a developmental paste were categorised to environmental testing and third harmonic measurement. The second category is the measurement AC Impedance and DC resistance of conductive ink during cure. During the curing of the pastes, the AC impedance and DC resistance were monitored. A LabVIEW program was developed to control the AC impedance analyser, DC resistance ohmmeter, and convection oven. Samples were measured whilst curing at different curing temperatures and for a range of particle loadings. Particle loading is the percentage of conductive filler against the rest of the chemical in the conductive paste. The last category was defect detection using the combination of electromechanical testing, a Scanning Electron Microscope (SEM) and an Infrared (IR) imaging technique. Printed carbon and silver were mechanically aged by bending the printed structure up to 100 k times. The results from the lifetime performance measurements on carbon, silver and the developmental paste showed the polymer resin behaviour in high humidity and high temperature environments. The increased oxidation rate due to the elevated temperatures affected the conductive particle of certain pastes. The third harmonic testing technique was able to detect failures on conductive film in the form of width reduction. The AC impedance measurement technique could indicate the final resistivity value. The AC impedance measurement was affected by the test frequency used while the ink is in liquid state. Correct test frequency setting will have less noise and less impedance value, vital in predicting the final cured resistance of the printed paste. The curing temperature affects the final cured resistance value while the particle loading affects the rate of curing of conductive film. The electrical measurement on mechanically aged samples showed that the carbon prints have its resistance readings below its initial value while the silver prints resistance increased. SEM images shows that the carbon print indicates no visual damage on the surface after 100 k bent cycle, while physical defects were observed in silver prints. The infrared measurements on carbon prints showed an increase in temperature while developments of heat patches were observed on silver prints. Difference in emissivity values of materials used provided the contrast effect which plays an important role in detecting defect using infrared imaging technique because. Third harmonic application to the printed electronics is new to this field. Normally, testing is done using environmental testing to determine the lifetime performance of the conductive film. This is effective however requires a lot of time and effort to produce a result. AC Impedance is used widely and the application can be seen on cured printed electronics. The application and measurement of AC impedance during cure and DC resistance measurement has indicated initial resistivity values. The measurement has further the effect of using AC impedance on different curing temperature and particle loading. The phase measurement as well has brought insight of degree of curing. The application of infra-red imaging technique to the mechanically aged device has produced a result that DC resistance and SEM imaging failed to detect. Normally DC resistance measurement was used as quality assessment tool but test shows on mechanically aged product failed to detect increase in resistance due to mechanical aging techqnique.

A gestão de clientes é um processo que envolve a tomada de decisões estratégicas, que influenciam a composição do portfólio de clientes da companhia, e operacionais, que afetam o relacionamento dos clientes com a empresa no dia a dia. O framework sugerido nesta tese propicia a integração dessas duas perspectivas, permitindo aos gestores alocarem melhor os recursos de marketing, por possibilitarem (a) o incremento da eficiência da carteira de clientes, a partir da sua otimização, e (b) a identificação dos clientes mais propensos a gerarem lucros futuros, com base na modelagem de customer lifetime value (CLV) desenvolvida. A abordagem de otimização do portfólio de clientes foi elaborada para auxiliar os gestores a definirem os segmentos que devem ser alvo dos investimentos de marketing e tem como objetivo indicar a composição da carteira de clientes que proporcionará a rentabilidade, a diversificação do risco e a lucratividade desejadas pelos acionistas. A abordagem sugerida é uma adaptação para o marketing da teoria financeira do portfólio. Foram incluídas restrições específicas para a área de gestão de clientes que asseguram a exequibilidade dos portfólios recomendados, tanto em relação à necessidade de aquisição de clientes ou de redução da participação dos segmentos na carteira, quanto em relação à manutenção da lucratividade da empresa. Ademais, foram incorporadas opções de estimação do retorno, tais como a inclusão da tendência à série com base na modelagem SUR, além de serem avaliadas a utilização de duas proxies para o risco, a variância e o Conditional Value at Risk. De acordo com o framework de gestão de clientes proposto, a implementação das decisões estratégicas é viabilizada a partir da integração da análise dos resultados obtidos pela otimização com a avaliação proporcionada pelo modelo de CLV sugerido. Este, além de englobar a evolução do comportamento do cliente ao longo do relacionamento da empresa, considera o retorno e a matriz de probabilidade de troca de segmento de maneira individualizada. A heterogeneidade da matriz de Markov foi alcançada a partir da combinação convexa da matriz de transição geral com a matriz personalizada de cada cliente, possibilitando, assim, a priorização de clientes pertencentes a um mesmo segmento. O framework sugerido foi aplicado na base de clientes de uma grande empresa que atua nacionalmente na indústria de serviços financeiros. Após a constatação de que os segmentos podem gerar diferentes retornos e representar distintos níveis de risco para a companhia, foi feita a comparação dos resultados dos portfólios recomendados com o realizado. Os portfólios sugeridos desempenharam melhor de maneira consistente em termos de lucratividade e de eficiência, medida a partir do sharpe ratio. Em relação ao modelo de CLV, os resultados foram comparados com os obtidos a partir do modelo de Pfeifer & Carraway (2000), utilizado como ponto de partida para o seu desenvolvimento. As modificações incorporadas, além de possibilitarem a individualização por cliente, aumentaram a precisão da previsão dos valores individuais e a qualidade do ordenamento, mantendo a capacidade de avaliação do valor da base. Para resumir, foi proposto um framework de gestão de clientes que inclui a avaliação do risco, possibilitando aos gestores uma visão holística do negócio e particular de cada cliente. / Customer management is a process that involves strategic decision-making, which influence the composition of the customer portfolio, and operational decision making, which affect the relationship of each customer with the company. The proposed framework provides the integration of the strategic and operational perspectives, empowering managers to better allocate marketing resources as it enables (a) the increase of the efficiency of the customer portfolio, through its optimization, and (b) the identification of the customers that are more likely to bring profit in the future, through the customer lifetime value (CLV) model developed. The customer portfolio optimization method was built to help managers to define the customer segments that should be the target of their marketing investments. Its purpose is to indicate the customer portfolio composition that will provide the return, profitability and risk diversification desired by shareholders. The suggested approach is an adaptation to marketing of financial portfolio theory. In this way, customer management specific constrains were included to ensure the applicability of the recommended portfolios in terms of either the necessity of acquiring new customers or reducing the importance of a given segment in the portfolio as well as in terms of maintaining the company’s profitability. Furthermore, options of estimating return were incorporated such as the inclusion of the trend in the time series based SUR modeling as well as the optimizations were evaluated considering two proxies for risk, variance and Conditional Value at Risk. According to the proposed framework, the implementation of the strategic decisions concerning the changes needed in the customer portfolio become possible through the integration of the results of the optimization with the estimation of the value of each customer provided by the CLV model developed. In this model, besides accounting for the evolution of the customer behavior throughout the duration of his relationship with the company, we also consider, for each customer, his individual return and his individual transition matrix. The heterogeneity of the Markov matrix was reached with a convex combination of the general transition matrix and the personalized matrix of each customer. It, therefore, enables managers to priorize customers of the same segment. The suggested framework was applied to the customer database of a large national company from the financial services industry. Once evidenced that the customer segments can generate different returns and can have different levels of risk for the company, we compared the results of the recommended with the current. The portfolios suggested by the optimization performed consistently better in terms of profitability and efficiency, measured through sharpe ratio. Concerning the CLV model developed, we compared the results with Pfeifer & Carraway (2000) model, which was used as the start point for our model. The improvements implemented not only allowed the estimation of CLV at the individual level, but also increased the precision of the predictions for the customer lifetime values and for the customer ranking, maintaining the quality of the customer equity forecast. To sum up, our proposed framework which includes risk assessment enables marketing managers to have a holistic vision of their customer portfolio and to drilldown into a particular vision of each customer.

A questão da produtividade das atividades de marketing é um antigo debate da disciplina, porém ganhou destaque nos anos 1990 a partir de uma conjuntura que propiciou questionamentos sobre a eficácia e eficiência dos departamentos de marketing das empresas. Argumentou-se que o marketing deveria ir além das métricas tradicionais e explicitar as consequências de marketing para todos os stakeholders. Nesse sentido, os conceitos de customer equity e customer lifetime value têm se mostrado promissores em fazer a conexão marketing-finanças a partir de uma série de estudos iniciados nos anos 1990. Desde então foram elaboradas inúmeras modelagens a fim de mensurar a contribuição das ações de marketing empiricamente. De fato, evidências apontam para uma real contribuição de marketing dentro da empresa, seja pelo valor da marca, da clientela, da satisfação, lealdade entre outros. No entanto, a diversidade de modelos pode significar um entrave na padronização de cálculo e não existem pesquisas que estudem o impacto do modelo escolhido nos resultados da estimação nem sua implicação para tomada de decisão. A presente tese visou investigar até que ponto existe modelo dependência na estimação do customer equity e customer lifetime value, já que os resultados impactam diretamente na estratégia a ser utilizada pela gerência e, consequentemente, no desempenho da empresa A partir da classificação dos modelos foram comparados aqueles chamados de determinísticos, um grupo de modelos com características semelhantes e implementação simplificada. Os resultados indicam que, no caso da utilização do customer equity para avaliação financeira da empresa a estimação apresenta grande variabilidade, o que indica modelo dependência e a necessidade de aprofundamento metodológico, principalmente em negócios com alta heterogeneidade de margens e retenções. O ciclo de vida do cliente, por sua vez apresenta modelo dependência no que diz respeito a estratégia gerencial para maximizar o mesmo, sendo a seguinte: nos modelos de projeção infinita a retenção de clientes apresenta “prêmio” crescente no ciclo de vida esperado, enquanto que margens de contribuição tem peso maior em ambientes com retenção menor que 50% aproximadamente. Modelos com projeção finita, no entanto, não apresentam essa característica, sendo a retenção sempre a causar maior impacto marginal no valor do ciclo de vida esperado. Esses resultados fornecem evidências para concluirmos que existe modelo dependência na estimação de tais métricas. / The marketing productivity is an old debate in the field and has gained substantial attention since 1990’s as a product of a context that raised questions about the real contribution of marketing, its efficiency and efficacy. Managers argued that marketing should demonstrate its effects beyond traditional metrics and making them available to all business’ stakeholders. In this sense, customer equity and customer lifetime value has been used to make the connection between marketing and finance by 1990’s until now. In fact, research has shown that marketing contributes to firm value by elevating brand equity, customer loyalty and satisfaction, customer equity, among others. However, the large number of customer equity and customer lifetime models can be a difficulty on finding standard methodology, and there are no studies that compare model’s impact on results. This study aims to investigate the dependency of the models in estimating customer equity and customer lifetime value and its consequence on strategy adoption by managers. Starting from a classification proposal, a group of deterministic models were chosen to be compared. Results show that using customer equity for valuation purposes is model dependent, since models present great estimation variability, especially in a heterogeneous context. Customer lifetime value does produce divergent strategies and has model dependency as well: on models of infinite projection retention rates have increasing effect on expected customer lifetime values, but margins are preferred until a retention threshold of fifty percent. This does not occur within a model with finite projection, which retention rates has always more impact than margins. The sum of this results offer evidence that, indeed, model dependency on estimating customer equity and customer lifetime value exists.

A Eletromigração (EM) nas interconexões de metal em um chip é um mecanismo crítico de falhas de confiabilidade em tecnologias de escala nanométrica. Os trabalhos na literatura que abordam os efeitos da EM geralmente estão preocupados com estes efeitos nas redes de distribuição de potência e nas interconexões entre as células. Este trabalho aborda o problema da EM em outro aspecto, no interior das células, e aborda especificamente o problema da eletromigração em interconexões de saída, Vdd e Vss dentro de uma célula padrão onde há poucos estudos na literatura que endereçam esse problema. Até onde sabe-se, há apenas dois trabalhos na literatura que falam sobre a EM no interior das células. (DOMAE; UEDA, 2001) encontrou buracos formados pela EM nas interconexões de um inversor CMOS e então propôs algumas ideias para reduzir a corrente nos segmentos de fio onde formaram-se buracos. O outro trabalho, (JAIN; JAIN, 2012), apenas cita que a EM no interior das células padrão deve ser verificada e a frequência segura das células em diferentes pontos de operação deve ser modelada. Nenhum trabalho da literatura analisou e/ou modelou os efeitos da EM nos sinais dentro das células. Desta forma, este é o primeiro trabalho a usar o posicionamento dos pinos para reduzir os efeitos da EM dentro das células. Nós modelamos a eletromigração no interior das células incorporando os efeitos de Joule heating e a divergência da corrente e este modelo é usado para analisar o tempo de vida de grandes circuitos integrados. Um algoritmo eficiente baseado em grafos é desenvolvido para acelerar a caracterização da EM no interior das células através do cálculos dos valores de corrente média e RMS. Os valores de corrente computados por esse algoritmo produzem um erro médio de 0.53% quando comparado com os valores dados por simulações SPICE. Um método para otimizar a posição dos pinos de saída, Vdd e Vss das células e consequentemente otimizar o tempo de vida do circuito usando pequenas modificações no leiaute é proposto. Para otimizar o TTF dos circuitos somente o arquivo LEF é alterado para evitar as posições de pino críticas, o leiaute da célula não é alterado. O tempo de vida do circuito pode ser melhorado em até 62.50% apenas evitando as posições de pino críticas da saída da célula, 78.54% e 89.89% evitando as posições críticas do pino de Vdd e Vss, respectivamente Quando as posições dos pinos de saída, Vdd e Vss são otimizadas juntas, o tempo de vida dos circuitos pode ser melhorado em até 80.95%. Além disso, nós também mostramos o maior e o menor tempo de vida sobre todos as posições candidatas de pinos para um conjunto de células, onde pode ser visto que o tempo de vida de uma célula pode ser melhorado em até 76 pelo posicionamento do pino de saída. Além disso, alguns exemplos são apresentados para explicar porque algumas células possuem uma melhora maior no TTF quando a posição do pino de saída é alterada. Mudanças para otimizar o leiaute das células são sugeridas para melhorar o tempo de vida das células que possuem uma melhora muito pequena no TTF através do posicionamento dos pinos. A nível de circuito, uma análise dos efeitos da EM é apresentada para as diferentes camadas de metal e para diferentes comprimentos de fios para os sinais (nets) que conectam as células. / Electromigration (EM) in on-chip metal interconnects is a critical reliability failure mechanism in nanometer-scale technologies. Usually works in the literature that address EM are concerned with power network EM and cell to cell interconnection EM. This work deals with another aspect of the EM problem, the cell-internal EM. This work specifically addresses the problem of electromigration on signal interconnects and on Vdd and Vss rails within a standard cell. Where there are few studies in the literature addressing this problem. To our best knowledge we just found two works in the literature that talk about the EM within a cell. (DOMAE; UEDA, 2001) found void formed due to electromigration in the interconnection portion in a CMOS inverter and then proposes some ideas to reduce the current through the wire segments where the voids were formed. The second work, (JAIN; JAIN, 2012), just cites that the standard-cell-internal-EM should be checked and the safe frequency of the cells at different operating points must be modeled. No previous work analyzed and/or modeled the EM effects on the signals inside the cells. In this way, our work is the first one to use the pin placement to reduce the EM effects inside of the cells. In this work, cell-internal EM is modeled incorporating Joule heating effects and current divergence and is used to analyze the lifetime of large benchmark circuits. An efficient graph-based algorithm is developed to speed up the characterization of cell-internal EM. This algorithm estimates the currents when the pin position is moved avoiding a new characterization for each pin position, producing an average error of just 0.53% compared to SPICE simulation. A method for optimizing the output, Vdd and Vss pin placement of the cells and consequently to optimize the circuit lifetime using minor layout modifications is proposed. To optimize the TTF of the circuits just the LEF file is changed avoiding the critical pin positions, the cell layout is not changed. The circuit lifetime could be improved up to 62.50% at the same area, delay, and power because changing the pin positions affects very marginally the routing. This lifetime improvement is achieved just avoiding the critical output pin positions of the cells, 78.54% avoiding the critical Vdd pin positions, 89.89% avoiding the critical Vss pin positions and up to 80.95% (from 1 year to 5.25 years) when output, Vdd, and Vss pin positions are all optimized simultaneously. We also show the largest and smallest lifetimes over all pin candidates for a set of cells, where the lifetime of a cell can be improved up to 76 by the output pin placement. Moreover, some examples are presented to explain why some cells have a larger TTF improvement when the output pin position is changed. Cell layout optimization changes are suggested to improve the lifetime of the cells that have a very small TTF improvement by pin placement. At circuit level, we present an analysis of the EM effects on different metal layers and different wire lengths for signal wires (nets) that connect cells.

El presente trabajo de titulo tiene como objetivo utilizar técnicas de data mining para determinar una metodología que permita estimar el lifetime value de los clientes de un supermercado mayorista. El comportamiento histórico de compra y variables de georeferenciación se utilizan para estimar cómo se comportará un cliente en el futuro. Este comportamiento se define como la variación porcentual del monto que desarrollará cada cliente. Para la estimación se utilizan técnicas probabilísticas y de data mining. En particular, se construyen cinco modelos basados en las siguientes técnicas: Pareto/NBD, Árbol de decisión y MLP. Posteriormente, se incluyen tres modelos ingenuos que permitan justificar el desarrollo de modelos sofisticados. El desempeño de los modelos indica que las técnicas de data mining, para el caso analizado, tienen mejores resultados en las estimaciones. Se concluye, que el modelo Pareto/NBD es conservador ya que tiende a asumir que un cliente mantendrá su comportamiento. No obstante, el modelo árbol de decisión es levemente más agresivo siendo el mejor a nivel de segmentos e individual con un error de precisión de un 55.8% a nivel de clientes y un 9.2% a nivel de segmentos. El resultado final consiste en la estimación del lifetime value a nivel de segmentos e individual de los clientes, utilizando el modelo árbol de decisión, que define un flujo monetario que se descuenta a lo largo de un periodo de tiempo. De esta manera se determina el lifetime value. Como trabajo futuro se propone definir un criterio que permita proyectar la tasa de crecimiento de los clientes a lo largo del tiempo.