Dissertations / Theses on the topic 'Stop loss'

Submitted by João Tóvolli (joaotovolli@hotmail.com) on 2016-02-27T17:09:13Z No. of bitstreams: 1 Dissertacao v11.docx: 1267937 bytes, checksum: c9a1fd7a1e77c02a843a4100d19fb25d (MD5)
Rejected by Renata de Souza Nascimento (renata.souza@fgv.br), reason: João, boa tarde De acordo com as normas da ABNT para que possamos aceitar seu trabalho, será necessário realizar alguns ajustes. Segue abaixo: - O arquivo deve estar em pdf - Na capa, retirar EESP que consta ao lado do nome da escola. - Na capa e contra capa deve estar SÃO PAULO 2016 (consta 2015) - Centralizar o título Resumo e Abstract - As numerações das páginas devem estar ao lado direito. Após alterações, submeter novamente o arquivo. Att on 2016-02-29T17:56:31Z (GMT)
Submitted by João Tóvolli (joaotovolli@hotmail.com) on 2016-02-29T18:16:54Z No. of bitstreams: 1 Dissertacao v12.pdf: 1025424 bytes, checksum: a080e366f58b59e6715da1ca07fe146b (MD5)
Approved for entry into archive by Renata de Souza Nascimento (renata.souza@fgv.br) on 2016-02-29T19:59:26Z (GMT) No. of bitstreams: 1 Dissertacao v12.pdf: 1025424 bytes, checksum: a080e366f58b59e6715da1ca07fe146b (MD5)
Made available in DSpace on 2016-03-01T12:16:39Z (GMT). No. of bitstreams: 1 Dissertacao v12.pdf: 1025424 bytes, checksum: a080e366f58b59e6715da1ca07fe146b (MD5) Previous issue date: 2016-02-02
This thesis proposes a trading framework to analyze the performance of Stop-Loss strategy in terms of value added. Based on the paper of Kaminsky e Lo (2014), the rule consists in switching from a high volatile asset to a risk free asset by a binary Stop-Loss rule. We found out an outstanding result, reducing volatility and an increasing coefficient return/volatility.
Nesta dissertação é proposta uma estrutura de trading para analisar o desempenho da estratégia de Stop-Loss em termos de ganho de valor. Fundamentada no paper de Kaminsky e Lo (2014), a regra consiste em alternar de um ativo de alta volatilidade para um ativo livre de risco baseado em uma regra binária de Stop-Loss. Foram encontrados resultados nominais positivos, redução de volatilidade e consequente aumento do coeficiente retorno/volatilidade.

Thesis (Ph. D.)--Electrical and Computer Engineering, Georgia Institute of Technology, 2007.
Tentzeris, Manos, Committee Member ; Wollack, Edward, Committee Member ; Cressler, John, Committee Member ; Papapolymerou, Ioannis, Committee Chair ; Laskar, Joy, Committee Co-Chair ; Ayazi, Farrokh, Committee Member.

Relatório de estágio do mestrado em Gestão, apresentado à Faculdade de Economia da Universidade de Coimbra, sob a orientação de Dinis Daniel dos Santos e António José Marques Mendes.
O presente relatório retrata o trabalho desenvolvido no âmbito do estágio curricular, realizado entre 19 de janeiro e 27 de maio de 2016, na SADIF – Investment Analytics. Num período em que os níveis de rendimento dos mercados financeiros apresentam, na sua generalidade, um fraco desempenho, verifica-se uma elevada concorrência na procura por desempenhos plausíveis, mas também, um grande recurso às tecnologias, para a melhor e maior difusão da informação e, com o objetivo de encontrar oportunidades de arbitragem. No investimento de longo prazo, torna-se essencial definir como deve ser feita a gestão, tendo por base um conjunto de orientações que prevaleçam sobre a mera intuição e que, por vezes, leva os investidores a tomar decisões pouco eficientes e que contrariam a racionalidade. Neste trabalho pretende-se encontrar evidências para a existência de regras predefinidas que apoiem a tomada de decisão na gestão de investimentos nos mercados financeiros, com o intuito de aumentar a rentabilidade. Ao longo do relatório o objetivo passa por apresentar uma reflexão sobre o desenvolvimento científico em torno dos investimentos financeiros e uma análise ao desempenho de algumas técnicas de gestão de carteiras, frequentemente reconhecidas enquanto ferramentas que, de um modo geral, oferecem melhores resultados nos investimentos. Para o estudo realizado foi utilizada a informação de um conjunto de carteiras de investimento, constituídas no âmbito da atividade da empresa acolhedora. As estatísticas apresentadas apontam para uma maior eficiência com a utilização do rebalanceamento trimestral e, também, com o recurso ao trailing-stop e ao stop-loss, enquanto ferramentas de apoio à decisão de vender antes do término do período de investimento definido.

The high prevalence and high costs of overweight or obesity in the United States, especially among veterans, presents a need for implementation of an effective weight loss program. Currently, Veteran Affairs hospitals use a weight loss and health promotion program called MOVE!, which has had problems with implementation and attendance. A new program called Aspiring to Lifelong Health in VA (ASPIRE) uses the Stop Light Diet (SLD) and the small change approach, and has been associated with significant weight loss and attainable implementation practices. This study reviews a 7 day food journal and weight change for 73 participants from both the MOVE! and ASPIRE programs within a three month time span. Particularly, change in consumption of foods based on the SLD categories and weight change were analyzed among participants in the two programs. Both programs resulted in significant weight loss (baseline to 3 months). The ASPIRE program was associated with an increase in "green foods," or fruits and vegetables, whereas the MOVE! program was not. When implementing a weight loss and health promotion program for the veteran population, a program using the concept of the SLD as well as providing a coach for the participants to set small and attainable dietary goals, using the small change approach, may help the participants increase their intake of fruits and vegetables, and decrease their intake of high calorie high fat foods in order to ultimately improve health and increase the chance for weight loss.

The diploma thesis deals with the possibilities of using technical analysis to trade on the financial markets, particularly in futures trading derived from stock market indexes. Specifically specifies selected trading system and then analyzes the results achieved by this system.

Mestrado em Finanças
Este estudo avalia a performance das mais comuns estratégias de Portfolio Insurance, baseando essa análise em simulações de blocos móveis de Bootstrap. Nesta análise consideramos não apenas as tradicionais medidas associadas à Teoria Média-variância, mas também outras medidas associadas ao Downside Risk, bem como classificações de dominância estocástica. Foram identificadas evidências que suportam que a estratégia CPPI 1 deve ser preferida em termos da sua dominância face às restantes. Contrariamente, a estratégia SLPI deverá ser preterida face a outras estratégias de Portfolio Insurance. Encontrámos igualmente evidências de que deverão ser escolhidas barreiras mínimas mais elevadas, com o objectivo de maximizar a utilidade da generalidade dos investidores. Consistentemente, e meramente em termos de performance, a estratégia CPPI 3 é aquela que apresenta resultados mais satisfatórios. Ao longo desta análise, tentamos proporcionar uma nova visão sobre as controversas estratégias de Portfolio Insurance, tentando tornar mais eficiente a decisão de futuros investidores.
This study evaluates the performance of the most common Portfolio Insurance Strategies based on a block-moving bootstrap simulation. We consider not only the traditional mean-variance approach, but also some measures of downside risk and stochastic dominance. We find that CPPI 1 should be preferred in terms of stochastic dominance. We also find that SLPI is constantly dominated by all the other strategies and a floor of 100% should be preferred to lower ones. Consistently, and purely in terms of performance analysis, CPPI 3 tends to outperform other strategies. During this analysis, we try to provide another insight into the controversy over Portfolio Insurance strategies, turning the decision-making process for future investors more efficient.

Thesis (MBA (Business Management))--University of Stellenbosch, 2009.
ENGLISH ABSTRACT: The purpose of this report is to investigate whether the returns of five different trading systems applied is able to outperform the return of a Buy & Hold (B&H) strategy when applied to the Johannesburg Stock Exchange/Financial Times Stock Exchange (JSE/FTSE) Top 40 Index future contract (ALSI). The study starts with an overview of theoretical and empirical studies regarding technical trading systems as well as the application of these technical trading systems in various strategy formats. Five common trading systems were selected for the test. They include the Volatility Channel, the Bollinger Channel Breakout, the Donchian Channel, the Dual Moving Average and the Triple Moving Average systems. The trading systems were applied in three different types of strategies. In the first test the systems were employed using randomly selected parameters to generate trading signals. In the second test the systems were optimised to select the parameters that would yield the most profitable returns over the test period. Finally in the third test a stop loss was added to the systems to investigate whether it would improve returns. In virtually all tests the systems outperformed the B&H approach. This was primarily due to the collapse of world financial markets in 2008 that caused the systems, which are all trend following by nature, to generate large returns. If it had not been for this event, the trend-following systems would all have underperformed the total return generated by the B&H strategy over the duration of the test period. The tests revealed that the selection of the parameters that generate the trade signals for the trading systems can drastically influence the profitability of a trading system. Furthermore the implementation of stop-loss strategies does not necessarily improve the return or drawdown that a system displays, as several of the systems were negatively influenced by the implementation of the stop-loss strategy.
AFRIKAANSE OPSOMMING: Die doel van hierdie verslag is om te ondersoek of die opbrengs van vyf verskillende verhandelingstelsels die opbrengs van die Koop-en-Hou-strategie kan klop soos toegepas op die JSE/FTSE Top 40 Indeks termynkontrak (ALSI). Die studie begin met ’n oorsig oor teoretiese en empiriese studies oor tegniese verhandelingstelsels, asook die toepassing van hierdie tegniese stelsels in verskeie strategiese formate. Vyf algemene verhandelingstelsels is gekies vir die ondersoek, naamlik die Volatiliteitskanaal (Volatility Channel), die Bollinger Kanaal Uitbreek (Bollinger Channel Breakout), die Donchian Kanaal (Donchian Channel), die Tweeledige Bewegende Gemiddelde (Dual Moving Average) en die Drieledige Bewegende Gemiddelde (Triple Moving Average). Die stelsels is op drie verskillende tipes stategieë toegepas. In die eerste toets was die stelsels geïmplementeer deur lukraak gekose parameters te gebruik om verhandelingseine voort te bring. In die tweede toets was die stelsels geoptimaliseer deur die parameters te kies wat die mees winsgewende opbrengs oor die toetsperiode sou voortbring. In die derde toets was ’n staakverlies (stop loss) geïmplementeer om te ondersoek of dit die opbrengs sou verbeter. Feitlik al die toetse het getoon dat die verhandelingstelsels die Koop-en-Hou-benadering geklop het. Aangesien al die stelsels die algemene tendens in die mark volg, het hulle hoë opbrengste getoon hoofsaaklik as gevolg van die beermark wat die wêreld se finansiële markte in 2008 gekenmerk het. As hierdie gebeurtenis nie plaasgevind het nie, sou hierdie stelsels swakker gevaar het as die Koop-en-Hou-strategie gedurende die tydperk van die toetsperiode. Die toetse het aangedui dat die keuse van die parameters wat verhandelingseine vir die stelsels gegenereer het, die winsgewendheid van ’n verhandelingstelsel drasties kan beïnvloed. Die implementering van ’n staakverlies- (stop-loss) strategie verbeter nie noodwendig die opbrengs van ’n stelsel nie, aangesien verskeie stelsels negatief beïnvloed was deur die staakverlies-strategie.

The purpose of this thesis is: (i) to produce an in-depth data analysis and computer-based simulations of the market environment to investigate whether financial stability is affected by the presence of High-Frequency investors; (ii) to verify how High-Frequency Trading and financial stability interact with each other under non-linear conditions; (iii) whether non-illicit behaviours can still lead to potentially destabilising effects; (iv) to provide quantitative support to the theses, either from the audit trail data or resulting from simulations. Simulations are provided to test whether High-Frequency Trading: (a) has an impact on market volatility, (b) leads to market splitting into two tiers; (c) takes the lion's share of arbitrage opportunities. Audit trail data is analysed to verify some hypotheses on the dynamics of the Flash Crash. The simulation on the impact of High-Frequency Trading on market volatility confirms that when markets are under stress, High-Frequency Trading may cause volatility to significantly increase. However, as the number of ultra-fast participants increases, this phenomenon tends to disappear and volatility realigns to its standard values. The market tiering simulation suggests that High-Frequency traders have some tendency to deal with each other, and that causes Low-Frequency traders also to deal with other slow traders, albeit at a lesser extent. This is also a kind of market instability. High-Frequency Trading potentially allows a few fast traders to grab all the arbitrage-led profits, so falsifying the Efficient Market Hypothesis. This phenomenon may disappear as more High-Frequency traders enter the competition, leading to declining profits. Yet, the whole matter seems a dispute for abnormal gains only between few sub-second traders. All simulations have been carefully designed to provide robust results: the behaviours simulated have been drawn from existing literature and the simplifying assumptions have been kept to a minimum. This maximises the reliability of the results and minimizes the potential of bias. Finally, from the data analysis, the impact of High-Frequency Trading on the Flash Crash seems significant; other sudden crashes occurred since, and more can be expected over the next future. Overall, it can be concluded that High-Frequency Trading shows some controversial aspects impacting on financial stability. The results are at a certain extent confirmed by the audit trail data analysis, although only indirectly, since the details allowing the match between High-Frequency traders and their behaviour are confidential and not publicly available Nevertheless, the findings about HFT-induced volatility, market segmentation and sub-optimal market efficiency, albeit not definitive, suggest that careful monitoring by regulators and policy-makers might be required.

Mestrado em Finanças
Este estudo realiza uma comparação entre as estratégias mais populares de Portfolio Insurance, através da Simulação de Monte Carlo. Este trabalho tem como objectivo definir a melhor estratégia através de diversas comparações e dar um contributo para resolver algumas divergências na literatura. A maioria das comparações realizadas anteriormente não têm em consideração todas as estratégias presentes neste estudo e esta análise pretende acrescentar algumas conclusões relevantes. As estratégias OBPI, CPPI e SLPI são avaliadas através dos momentos da distribuição, rácios de desempenho (Sharpe ratio, Sortino ratio, Omega ratio e Upside Potential ratio) e dominâncias estocásticas nas diversas condições de mercado representadas pelo activo subjacente que segue um movimento Browniano geométrico. De forma a ter uma compreensão da realidade dos mercados financeiros, as estratégias também são aplicadas a três dos maiores índices de acções. Concluímos que as estratégias CPPI 1 e SLPI devem ser preferidas em todos os cenários devido aos elevados rácios de desempenho, elevadas rendibilidades esperadas e a outras medidas. A escolha entre as duas estratégias é feita com base nas preferências do investidor ou gestor, mas também concluímos que a estratégia CPPI 1 domina estocásticamente, a segunda e terceira ordem, todas as restantes estratégias em cenários de mercado bear. De acordo com os resultados obtidos podemos afirmar que um floor de 100% deve ser escolhido devido aos resultados dos rácios de desempenho, rendibilidades esperadas e outras medidas. Esta comparação permite melhorar a eficiência da tomada de decisão de um investidor ou gestor num investimento de Portfolio Insurance.
This study makes a comparison between the most popular strategies of Portfolio Insurance based on Monte Carlo simulation. This work aims to define the best strategy at comparing different strategies and provide a contribution to solving some divergences in literature. Most of the previous comparisons do not take into consideration all the strategies discussed in this study and this analysis intends to add some relevant findings. The OBPI, CPPI and SLPI strategies are evaluated in terms of moments of the distribution, performance ratios (Sharpe ratio, Sortino ratio, Omega ratio and Upside Potential ratio) and stochastic dominance in different market conditions represented by an underlying asset that follows a geometric Brownian motion. In order to have a perception of a real situation in financial markets, the strategies are later also applied to three major stock indices. We find that CPPI 1 and SLPI strategies should be preferred in all scenarios according to the higher performance ratios, the higher expected returns and other measures. The choice between them is based on the preferences of the investor or manager, but we also find that the CPPI 1 strategy stochastically dominates, on second and third order, the others strategies in bear market scenarios. From our results we can state that a value of 100% for the floor should be preferred in terms of performance ratios, expected returns and other measures. This comparison allows improving the efficiency of decision making of an investor or manager in a Portfolio Insurance investment.

Práca si kladie za cieľ vytvorenie obchodného systému pre intradenné obchodovanie US akciových indexov. Autor sa v teoretických východiskách zameriava na vysvetlenie základných pojmov obchodovania US indexov na intradennej báze a obchodovania všeobecne. V ďalšej kapitole je popísaná súčastná situácia – tvorba obchodného plánu a vysvetlené principy na ktorých je plán založený. V návrhovej časti je predstavený kompletný systém pre intradenné obchodovanie US akciových indexov.

Cette thèse utilise des outils théoriques de la finance, de la théorie de la décision et de l'apprentissage automatique, pour améliorer la conception, la tarification et la couverture des contrats d'assurance. Le chapitre 3 de cette thèse développe des formules de tarification sous forme fermée pour une classe de contrats d'assurance vie participatifs, sur la base de la factorisation matricielle de Wiener-Hopf, et prend en compte plusieurs types de risque, tels que les risques de crédit, de marché et économiques. La méthode de tarification se révèle précise et efficace. Les stratégies de couverture dynamique et semi-statique sont introduites pour aider les compagnies d'assurance à réduire leur risque lié à l'émission de contrats participatifs. Le chapitre 4 traite de la conception optimale de contrats lorsque l'assuré possède une aversion au risque du troisième degré. Les résultats exhibent une forme de contrat optimale pour les agents averses au risque comme pour ceux appréciant le risque dans différents contextes. Le chapitre 5 développe un modèle stochastique amplificateur degradient fréquence/sévérité qui améliore les modèles de fréquence et de sévérité importants et populaires que sont les modèles GLM et GAM. Ce nouveau modèle hérite pleinement des avantages de l'algorithme de renforcement du gradient, dépassant ainsi les formes linéaires ou additives restrictives des modèles GLM et GAM, avec apprentissage de la structure du modèle à partir des données. En outre, ce modèle peut également rendre compte de la dépendance non linéaire existant entre fréquence et sévérité des sinistres
This thesis makes use of some theoretical tools in finance, decision theory, machine learning, to improve the design, pricing and hedging of insurance contracts. Chapter 3 develops closed-form pricing formulas for participating life insurance contracts, based on matrix Wiener-Hopf factorization, where multiple risk sources, such as credit, market, and economic risks, are considered. The pricing method proves to be accurate and efficient. The dynamic and semi-static hedging strategies are introduced to assist insurance company to reduce risk exposure arising from the issue of participating contracts. Chapter 4 discusses the optimal contract design when the insured is third degree risk averse. The results showthat dual limited stop-loss, change-loss, dual change-loss, and stop-loss can be optimal contracts favord by both of risk averters and risk lovers in different settings. Chapter 5 develops a stochastic gradient boosting frequency-severity model, which improves the important and popular GLM and GAM frequency-severity models. This model fully inherits advantages ofgradient boosting algorithm, overcoming the restrictive linear or additive forms of the GLM and GAM frequency-severity models, through learning the model structure from data. Further, our model can also capture the flexible nonlinear dependence between claim frequency and severity

Abstract The thesis deals with the priváte banking in the Czech Republic and in Liechtenstein. The goal is to analyze concrete services that the client can expect in both of these countries. Large topic in this thesis is the lombard credit, which is quite a typical product in western countries but not that known among Czech investors. As well as the thesis describes all the mechanisms around it.

In this thesis, we address two major problems in Generative Adversarial Networks (GAN), an important sub-field in deep learning. The first problem that we address is the instability in the training process that happens in many real-world problems and the second problem that we address is the lack of a good evaluation metric for the performance of GAN algorithms. To understand and address the first problem, three approaches are developed. Namely, we introduce randomness to the training process; we investigate various normalization methods; most importantly we develop a better parameter initialization strategy to help stabilize training. In the randomness techniques part of the thesis, we developed two randomness approaches, namely the addition of gradient noise and the batch random flipping of the results from the discrimination section of a GAN. In the normalization part of the thesis, we compared the performances of the z-score transform, the min-max normalization, affine transformations and batch normalization. In the most novel and important part of this thesis, we developed techniques to initialize the GAN generator section with parameters that can produce a uniform distribution on the range of the training data. As far as we are aware, this seemingly simple idea has not yet appeared in the extant literature, and the empirical results we obtain on 2-dimensional synthetic data show marked improvement. As to better evaluation metrics, we demonstrate a simple yet effective way to evaluate the effectiveness of the generator using a novel "overlap loss".

Tesis para optar al grado de Magíster en Economía
Autor no autoriza el acceso a texto completo de su documento
Este trabajo complementa las ideas de Agosin y Huaita (2012) en “Overreaction in Capital Flows to Emerging Markets: Booms and Sudden Stops”, pero introduce la distinción entre flujos IED y no-IED, utilizando los flujos non-IED de la cuenta financiera de la balanza de pagos (excluyendo movimientos de reservas) como medida de los flujos de capital netos en lugar de la cuenta financiera total. La hipótesis que sostiene este documento es que son grandes flujos distintos de la inversión extranjera directa los que tienen un efecto importante en la predicción de Sudden Stops de capitales. Complementando estas ideas, se agrega al análisis la descomposición de los flujos no- IED en dos grandes componentes principales - flujos de cartera y derivados financieros y otras inversiones - con el objetivo de testear su impacto tanto sobre episodios de Sudden Stops como sobre episodios de Booms. Los resultados arrojan que Booms liderados por flujos no-IED son el principal causante de Sudden Stops. Para Booms de capitales se obtiene el mismo patrón. Sin embargo, se observa una relación entre grandes flujos de inversión extranjera directa y estos episodios, aunque esta relación es más débil que el resto de las partidas de la cuenta financiera.

Made available in DSpace on 2010-04-20T21:00:36Z (GMT). No. of bitstreams: 4 Daniel Guedini.pdf.jpg: 2593 bytes, checksum: b9f66c6e9af5d94ff028f0092fbf9114 (MD5) Daniel Guedini.pdf.txt: 116251 bytes, checksum: 11ea61b001a20cef77fcf9db8c74c28d (MD5) license.txt: 4712 bytes, checksum: 4dea6f7333914d9740702a2deb2db217 (MD5) Daniel Guedini.pdf: 1363558 bytes, checksum: 4e6b49114a354eeaa85c0d6710a75c84 (MD5) Previous issue date: 2010-01-28T00:00:00Z
Diante do inédito momento vivido pela economia brasileira e, especialmente, pela bolsa de valores nacional, principalmente após a obtenção do grau de investimento pelo Brasil, este trabalho aborda um tema que ganhou um enorme espaço na mídia atual que é a análise técnica. A partir de uma amostra de 37 ações listadas na Bolsa de Valores de São Paulo no período compreendido entre janeiro de 1999 e agosto de 2009, este trabalho examina se a análise técnica agrega valor 'as decisões de investimentos. Através da elaboração de intervalos de confiança, construídos através da técnica de Bootstrap de inferência amostral, e consistentes com a hipótese nula de eficiência de mercado na sua forma fraca, foram testados 4 sistemas técnicos de trading. Mais especificamente, obteve-se os resultados de cada sistema aplicado às series originais dos ativos. Então, comparou-se esses resultados com a média dos resultados obtidos quando os mesmos sistemas foram aplicados a 1000 séries simuladas, segundo um random walk, de cada ativo. Caso os mercados sejam eficientes em sua forma fraca, não haveria nenhuma razão para se encontrar estratégias com retornos positivos, baseando-se apenas nos valores históricos dos ativos. Ou seja, não haveria razão para os resultados das séries originais serem maiores que os das séries simuladas. Os resultados empíricos encontrados sugeriram que os sistemas testados não foram capazes de antecipar o futuro utilizando-se apenas de dados passados. Porém, alguns deles geraram retornos expressivos e só foram superados pelas séries simuladas em aproximadamente 25% da amostra, indicando que a análise técnica tem sim seu valor.
Faced with unprecedented time lived by Brazilian`s economy and, especially, the national stock exchange, mainly after obtaining the investment grade for Brazil, this paper addresses a theme that has deserved a huge space in the mainstream media that is technical analysis. From a sample of 37 stocks listed on the Stock Exchange of Sao Paulo in the period between January 1999 and August 2009, this paper examines if the technical analysis may or may not add value to investment decisions. Through the development of confidence intervals, constructed using the technique of Bootstrap sample inference, and consistent with the null hypothesis of market efficiency in its weak form, we tested 4 technical systems of trading. More specifically, we obtained the results of each system applied to the original series of the assets. Then we compared these results with the average of the results obtained when the same systems were applied to 1000 simulated series, according to a random walk, of each asset. If markets are efficient in its weak form, there would be no reason to find strategies with positive returns based only on historical values of assets. That is, there would be no reason for the results of the original series to be larger than those of the simulated series. The empirical results found here suggested that the systems tested were unable to anticipate the future using only past data. However, some of them have generated significant returns and were surpassed only by the series simulated in approximately 25% of the sample, indicating that technical analysis does have value.

Les systèmes de traitement biologique de l'azote, qui utilisent les processus de nitrification et de dénitrification, sont susceptibles d'émettre un gaz à effet de serre, l'oxyde nitreux (N2O). Le but de notre étude était d'estimer et d'identifier les processus et les facteurs responsables de ces émissions, provenant du traitement de l'azote sur les stations d'épuration de l'agglomération parisienne et d'en estimer l'impact sur les émissions globales de N2O au niveau du bassin de la Seine. Des expériences en laboratoire ont permis la simulation des différents traitements utilisés sur l'agglomération parisienne - boues activées secondaires et cultures fixées en traitement tertiaire - et la mesure, en conditions contrôlées, des émissions de N2O. Nous avons montré que les flux de N2O représentent entre 0.1 à 0.8 % de la charge en azote traitée suivant l'oxygénation et les doses de méthanol ajoutées. Les flux de N2O les plus importants sont observés pour une oxygénation autour de 1 mgO2 L-1 lors de la nitrification; autour de 0.3 mgO2 L-1 et pour un ajout de méthanol en traitement tertiaire, ne permettant que 60 % de la réduction totale des nitrates, lors de la dénitrification. Des essais expérimentaux utilisant différents inhibiteurs spécifiques, ont permis d'identifier les deux processus majeurs responsables de ces émissions: la nitrification-dénitrifiante autotrophe et la dénitrification hétérotrophe. Le traitement de l'azote, sur les stations de l'agglomération parisienne représenterait actuellement des émissions de N2O de l'ordre de 60-120 kgN-N2O j-1 qui augmenteront avec la mise en place des nouveaux traitements de l'azote pour l'horizon 2005-2008 à 320-480 kgN j-1, et pour l'horizon 2012-2015 à 370-750 kgN j-1. Les niveaux les plus bas de ces émissions pourront être atteints si le traitement de l'azote se fait en nitrification avec une oxygénation supérieure à 2 mgO2 L-1 et en dénitrification avec des conditions d'anoxie totales et un ajout de méthanol permettant 100 % de la réduction de la charge en nitrate. Nos résultats montrent également que les émissions de N2O, se produisant actuellement dans la basse Seine recevant les rejets azotés de l'agglomération parisienne augmenteront notablement en étant transférées vers les STEPs. Cependant, les émissions de N2O en STEPs ne représenteraient que 1 à 10 % des émissions provenant des sols agricoles.

A limiting factor on DSSC efficiency is the lower fraction of the solar spectrum that is absorbed by the dye molecules developed to this point. Dye molecules that function well in DSSCs tend to have poor or no absorption to the red of 750 nm. Extending this absorption to the red by 100 nm, without losing efficiency in other ways, would result in a significant improvement in photocurrent. This challenge has proven difficult, in large part because of one slow reaction in the electron transfer cycle of DSSCs, the regeneration reaction. Better understanding of this reaction is thus critical. The kinetics of regeneration is understudied relative to the other processes in DSSCs, this is in part because the regeneration reaction produces no, as yet detected, measurable electrical signal. It must be studied by more difficult transient absorbance (TA) techniques. The first step of this thesis focuses on isolating a reliable transient signal that reflects the regeneration reaction. This is made by upgrading the conventional TA system to also acquire transient electrical (TE) signals simultaneously (TA-TE). The system is used to characterize dye-sensitized solar cells (DSSCs) under 1 sun illumination whilst the cells are fully operational and their stability is monitored. The second step of the work consists of the development of a methodology and a kinetic model which uses the isolated regeneration signal and a range of complimentary measurements on operating cells, to determine the quantum yield and the associated intrinsic rate constants and orders of the regeneration reaction. This enabled understanding of the regeneration mechanism and its optional rate limiting steps. Finally, the use of steady state photoinduced absorption (SSPA), as a complementary or alternative tool to assess regeneration, is also questioned. SSPA is compared with the regeneration TA –TE and charge extraction measurements.

Submitted by Karlla Delalibera Chagas null (karlladelalibera@gmail.com) on 2018-05-14T12:25:13Z No. of bitstreams: 1 dissertação - Karlla Delalibera.pdf: 2936984 bytes, checksum: 3d99ddd54b4c7d3230e5de9070915594 (MD5)
Approved for entry into archive by Claudia Adriana Spindola null (claudia@fct.unesp.br) on 2018-05-14T12:53:09Z (GMT) No. of bitstreams: 1 chagas_kd_me_prud.pdf: 2936984 bytes, checksum: 3d99ddd54b4c7d3230e5de9070915594 (MD5)
Made available in DSpace on 2018-05-14T12:53:09Z (GMT). No. of bitstreams: 1 chagas_kd_me_prud.pdf: 2936984 bytes, checksum: 3d99ddd54b4c7d3230e5de9070915594 (MD5) Previous issue date: 2018-04-16
Pró-Reitoria de Pós-Graduação (PROPG UNESP)
Neste trabalho iremos realizar uma abordagem sobre a modelagem de dados que advém de um teste acelerado. Consideraremos o caso em que a carga de estresse aplicada foi do tipo "step-stress". Para a modelagem, utilizaremos os modelos step-stress simples e múltiplo sob censura tipo II e censura progressiva tipo II, e iremos supor que os tempos de vida dos itens em teste seguem uma distribuição Gama. Além disso, também será utilizado o modelo step-stress simples sob censura tipo II considerando a presença de riscos competitivos. Será realizada uma abordagem clássica, por meio do método de máxima verossimilhança e uma abordagem Bayesiana usando prioris não-informativas, para estimar os parâmetros dos modelos. Temos como objetivo realizar a comparação dessas duas abordagens por meio de simulações para diferentes tamanhos amostrais e utilizando diferentes funções de perda (Erro Quadrático, Linex, Entropia), e através de estatísticas verificaremos qual desses métodos se aproxima mais dos verdadeiros valores dos parâmetros.
In this work, we will perform an approach to data modeling that comes from an accelerated test. We will consider the case where the stress load applied was of the step-stress type. For the modeling, we will use the simple and multiple step-stress models under censorship type II and progressive censorship type II, and we will assume that the lifetimes of the items under test follow a Gamma distribution. In addition, the simple step-stress model under censorship type II will also be used considering the presence of competitive risks. A classical approach will be performed, using the maximum likelihood method and a Bayesian approach using non-informative prioris, to estimate the parameters of the models. We aim to compare these two approaches by simulations for different sample sizes and using different loss functions (Quadratic Error, Linex, Entropy), and through statistics, we will check which of these approaches is closer to the true values of the parameters.

There is increasing evidence that aspects of human eating may be 'addictive'. Much of the existing literature examining this focuses on specific foods or individual pathologies of 'addiction'. Qualitative research methods, in particular content analysis, offers a rich opportunity to better understand 'addictive' aspects of human eating through the stories shared by those who have experienced compulsive eating. This research examines two different organizations that use storytelling as a tool for changing eating behaviours. Overeaters Anonymous (OA) and Weight Watchers (WW) began in the United States in the early 1960s as mutual support groups designed to help members with problems of compulsive overeating. This research examines: 1) the ways that OA and WW addressed 'addictions' with food when they were first formed; 2) how each organization has changed over time; and 3) the ways that identities are constructed through the telling of stories within each programme. For this thesis, I used historical analysis, in-person and online participant observation, and content analysis as research methods. I found that while OA and WW once similarly addressed 'addictions' with food, this is no longer the case. WW no longer understands their members as fundamentally different from others in the ways that they eat, and OA now welcomes members with a variety of different compulsive eating behaviours. However, similar themes regarding identity emerged from the OA and WW stories used in this research. OA and WW members describe that they acquire different belief systems regarding health and how to eat in each programme. In comparing these two organizations, how each has changed over time and the shifts in identity described by OA and WW members, this research identifies aspects of everyday living that members of both groups similarly highlight as important factors influencing compulsive eating. Based on these findings, I then assess to what extent a new framework for understanding 'addiction' may be needed, and conclude by suggesting further areas of research that would be suitable for constructing this.

INTRODUCCIÓ: Els medicaments que ajuden a deixar de fumar han demostrat ser segurs i eficaços (el seu ús pot evitar moltes morts). Les principals guies recomanen el seu ús en el 100% dels intents d’abandonament. En la Unió Europea s’usen en l’11% dels intents, a Espanya en el 3%. OBJECTIUS: conèixer l’ús d’aquests medicaments, l’ús del tabac, les creences respecte a aquestes medicaments i el paper del metge en les recomanacions sobre aquestes medicaments. Analitzar aquests objectius respecte a l’estatus socioeconòmic. MATERIAL I MÈTODE: Estudi descriptiu transversal. Es va elaborar un qüestionari basat en la teoria de l’acció raonada, teoria que permet predir la conducta, sota l’ús dels medicaments, a partir de les creences que es tinguin sobre ells. Es distribueixen el qüestionaris entre els usuaris de 3 oficines de farmàcia. RESULTATS: L’ús d’aquests medicaments és molt inferior al recomanat, la situació és pitjor en el cas dels medicaments que precisen recepta. Augmenta l’ús del tabac al disminuir el estatuts. Més de la meitat dels enquestats consideren que aquests medicaments son innecessària, només un de cada quatre considera que son eficaços, hi ha una gran desconeixement sobre la seguretat. Tres de cada quatre fumadors manifesten que el seu metge mai els hi ha recomanat aquests medicaments. DISCUSSIÓ: Atès que les creences que es tinguin sobre els medicaments determinen el seu ús i no és d’estranyar l’escàs ús ateses les creences errònies detectades. Possiblement les creences sobre la necessitat d’ús condicionin altres tipus de creences (de seguretat o d’eficàcia) i estan relacionades amb la idea sobre si el tabaquisme és un hàbit -en aquest cas la solució depèn de la força de voluntat- o una malaltia -susceptible de ser tractada amb medicaments-.
INTRODUCCIÓN: Los medicamentos que ayudan a dejar de fumar son seguros y eficaces (su uso puede evitar muchas muertes). Las guías recomiendan su uso en el 100% de los intentos de abandono. En la Unión Europea se usan en el 11% de los intentos y en España en el 3%. OBJETIVOS: Conocer el uso de estos medicamentos, el uso del tabaco, las creencias respecto a estos medicamentos y el papel del médico en la recomendación de los mismos. Analizar los objetivos anteriores respecto al estatus socioeconómico. MATERIAL Y MÉTODO: Estudio descriptivo transversal. Se elaboró un cuestionario basado en la teoría de la acción razonada, teoría que permite predecir la conducta, bajo uso de los medicamentos, a partir de las creencias que se tengan sobres los mismos. Se distribuyen los cuestionarios entre los usuarios de 3 oficinas de farmacia. RESULTADOS: El uso de estos medicamentos es muy inferior al recomendado y la situación es peor en los medicamentos que precisan receta. Aumenta el uso del tabaco al disminuir el estatus. Más de los encuestados consideran que estos medicamentos son innecesarios, solo uno de cada cuatro considera que son eficaces y hay un gran desconocimiento sobre la seguridad. Tres de cada cuatro fumadores manifiestan que el médico nunca les ha recomendado estos medicamentos. DISCUSIÓN: Dado que las creencias que se tengan sobre los medicamentos determinan su uso, no es de extrañar el escaso uso dadas las creencias erróneas detectadas. Posiblemente las creencias sobre sobre la necesidad de uso condicionen otros tipos de creencias (de seguridad o de eficacia) y están relacionadas con la idea sobre si el tabaquismo es un hábito -cuya solución depende de la fuerza de voluntad- o una enfermedad -susceptible de ser tratada con medicamentos-.
INTRODUCTION: Medicines that help quit smoking have been shown to be safe and effective (their use can prevent many deaths). The main guidelines recommend its use in 100% of quit attempts. In the European Union they are used in 11% of the attempts and in Spain in 3%. OBJECTIVES: To know the use of these medications, the use of tobacco, the beliefs regarding these medications and the role of the doctor in recommending them. Analyze the previous objectives regarding socioeconomic status. MATERIAL AND METHOD: Cross-sectional descriptive study. A questionnaire was developed based on the theory of reasoned action, a theory that allows predicting behavior, under the use of drugs, based on the beliefs held about them. The questionnaires are distributed among users from 3 pharmacy offices. RESULTS: The use of these drugs is much lower than recommended and the situation is worse in the case of drugs that require a prescription. Tobacco use increases by decreasing status. More than half of those surveyed consider that these drugs are unnecessary, only one in four considers them effective and there is a great lack of knowledge about their safety. Three out of four smokers say that their doctor has never recommended these drugs. DISCUSSION: Since beliefs about drugs determine their use and it is not surprising how little use is given due to the erroneous beliefs detected. Possibly beliefs about the need for use condition influence other types of beliefs (safety or efficacy) and are related to the idea of whether smoking is a habit -whose solution depends on willpower- or a disease -susceptible to be treated with medications-.

Cette étude avait pour objectif premier une caractérisation biomécanique du simple pas chez le sujet « normal ». À cet effet, on a considéré l’effet de la vitesse, ce qui a notamment permis de tester l’invariance de l’égalité entre la perturbation appliquée au corps lors de son accélération et la contre-perturbation lors de son freinage (CPA) permettant le retour au repos. Ainsi , le rôle des CPA se trouve explicité. Dans une seconde série expérimentale, on a considéré l’effet du frottement à l’arrêt du mouvement. L’objet de cette série était de caractériser la modification des caractéristiques biomécaniques chez le sujet « normal ». La caractérisation n’a pas soulevé de problèmes majeurs, l’effet de la vitesse se confirmant pour les deux coefficients de frottement (COF) considérés ( Téflon et Carrelage ). En revanche, les différences entre caractéristiques biomécaniques paraissent ne pas être significatives quand le COF diffère. Néanmoins, si l’on considère les résultats individuels, on constate que la durée des CPA tend à être systématiquement supérieure pour le COF le plus faible (Téflon), alors que le pic d’amplitude est systématiquement inférieur . Ces résultats incitent à approfondir l’étude de l’effet du frottement sur le simple pas. En conclusion , le simple pas semble être un paradigme permettant d’obtenir des résultats robustes et facilement utilisables pour étudier des personnes vieillissantes ou handicapées
This study aimed to biomechanical characterization of a simple step among normal subjects. To this end, we considered the effect of speed, which has enabled to test the invariance of equality between the disturbance applied to the body during its acceleration and counter- perturbation during its breaking phase (CPA) to return to initial position. Also, the role of CPA is explained.In a second experimental series, we considered the effect of friction at the stop point of movement .The purpose of this series was to characterization of biomechanical characteristics modifications among normal subjects. So, the characterization did not raise any major problems, the effects of velocity is confirmed for both COF considered (Teflon and Tiles). How ever, the differences between biomechanical characteristics are not significant when the COF is different, but by considering the individual results, we have found that the duration of the CPA tends to be systematically higher for Teflon, contrary to the peak amplitude results.These results encourage further study of the effect of friction on the single step.In conclusion, the simple step seems to be a paradigm for obtaining robust results and easily usable to study old or handicapped peoples

Chemotherapy is a general treatment for most breast tumors and depending on breast cancer subtype combination with targeted anti-hormonal or anti-HER2 therapy is conducted. Taxane-based regimens constitute the most common therapeutic option in patients with triple negative breast cancer (TNBC), but resistance often develops followed by metastatic disease and mortality. Aiming to reveal the mechanisms underlying taxane resistance, we used breast cancer patient-derived xenografts (PDX). Mimicking clinical behavior, triple-negative breast tumors (TNBCs) from PDX models were more sensitive to docetaxel than luminal tumors, but they progressively acquired resistance upon continuous drug administration. We analyzed cancer stem cell population dynamics in paired sensitive and resistant TNBC PDX models, but also epigenomics, genomics and transcriptomics profiles to elucidate the unknown molecular mechanisms responsible for this process. Mechanistically, we found that a CD49f+ chemoresistant population with tumor-initiating ability is present in sensitive tumors and expands during the acquisition of drug resistance. In the absence of the drug, the resistant CD49f+ population shrinks and taxane sensitivity is restored. We describe a transcriptional signature of resistance, predictive of recurrent disease after chemotherapy in TNBC. Together, these findings identify a CD49f+ population enriched in tumor-initiating ability and chemoresistance properties and evidence a drug holiday effect on the acquired resistance to docetaxel in triple-negative breast cancer. Exome sequencing was also performed in the matched sensitive and docetaxel resistant TNBC PDX models. Multiple mutations, small insertion/deletions and copy number variation (CNV) changes were detected in the original human metastatic samples, and most of them were maintained in serially passaged TNBC PDX models even after long term treatment with docetaxel, with very few changes being detected between paired sensitive and resistant tumors. However, we identified a chromosomal amplification of chr12p arm present in the metastatic sample and chemoresistant PDXs of one BRCA1 mutant models which was absent in docetaxel sensitive PDXs. Increased expression levels of genes located at chr12p correlated with amplificacion in chemoresistant tumors. Clinical data from TCGA and METABRIC studies confirmed that chr12p amplification was associated with a small subset of TNBC/basal-like breast cancer patients with increased gene expression of genes from that region and poor survival after chemotherapy. Our findings suggest that there is a subset of TNBC/basal-like breast cancer harbouring chr12p amplification that may be associated with resistance to docetaxel treatment in clinical setting. Genome-wide DNA methylation and gene expression analysis have been performed in preclinical breast cancer patient-derived xenograft (PDX) models with primary and acquired chemoresistance. These analyses revealed that DNA methylation patterns from breast cancer PDX are closer to breast cancer clinical samples than breast cancer cell lines (BCCLs) and they maintain subtype specific methylation patterns. Triple negative PDX tumors show very stable methylation patterns accompanying chemoresistance acquisition but some critical genes/pathways were unraveled as differentially methylated. Transcriptomically, TNBC PDX tumors accumulate gene expression changes during chemoresistance acquisition with some common pathways between different triple negative PDX models. Integrative analysis reveals correlation of some differentially methylated genes as differentially expressed in resistant triple negative breast cancer PDX tumors. These findings identify a set of promising pathways that may contribute to the acquisition of chemoresistance in TNBC patients. Receptor activator of NF-κB (RANK) is expressed in human breast tumors and has been associated with aggressive breast cancers. In this study we used breast cancer PDX models to investigate the functional role of RANK and its ligand (RANKL) in clinical human breast cancer. RANK expression in human breast cancer is more frequent in ER/PR-negative than in hormone receptor positive tumors and that is maintained in breast cancer PDX models. RANKL is generally poorly expressed in human breast cancer, but some RANKL-positive breast cancer PDX models were identified. Selection of RANK/RANKL-positive breast cancer PDX models expressing different levels of RANK or RANKL was done in order to functionally study the role of RANK signalling in human breast cancer. In vitro RANKL treatment on breast cancer cells isolated from RANK/RANKL-positive PDX models show modulation of NF-κB pathway, even in tumors where RANK expression was very low.
La quimioterapia es un tratamiento general usado comúnmente en cáncer de mama, sobretodo en tumores del subtipo triple negativo (TNBC). Éstos responden inicialmente muy bien, pero con el tiempo suelen dar lugar a recidivas o metástasis quimioresistentes. Empleando modelos preclínicos de cáncer de mama hemos estudiado los mecanismos subyacentes de la adquisición de resistencia a taxanos mediante análisis de poblaciones celulares, así como de análisis epigenómicos, genómicos y transcriptómicos Con el estudio de dinámicas poblacionales descubrimos la existencia en tumores TNBC inicialmente sensibles la existencia de una población CD49f+ con capacidad de iniciación tumoral que se expande durante la adquisición de quimioresistencia. Esta población revierte en ausencia de droga y con ella la quimioresistencia. Además, describimos una firma transcripcional de resistencia, predictivo de recidiva de la enfermedad después de la quimioterapia de TNBC. Los análisis genómicos revelaron la existencia de mútiples mutaciones y aberraciones cromosómicas en las muestras de pacientes que se mantienen de forma estable en los modelos preclínicos, aún con la adquisición de quimioresistencia. Únicamente se detectó una amplificación del cromosoma 12 en tumores resistentes no presente en sensibles. Esta amplificación se asocia en muestras clínicas a un subgrupo de tumores TNBC con peor supervivencia que el resto, pudiendo ser quimioresistentes. Los análisis epigenómicos revelaron la utilidad de estos modelos preclínicos al ser muy cercanos a muestras clínicas y mantener los patrones asociados a cada subtipo de cáncer de mama. Estos patrones son tan estables que prácticamente no varían con la adquisición de quimioresistencia, sin embargo algunos genes y rutas concretos sí lo hacen. Los análisis genómicos mostraron algunas vías comunes alteradas y un análisis integrador de los datos de metilación y expresión mostraron genes y rutas afectadas que presentaban metilación y expresión diferencial con la adquisición de resistencia.

Pour comprendre l’étiologie des chutes chez la personne âgée, les études menées dans le domaine de la biomécanique se sont principalement attachées à évaluer les capacités d’équilibre en s’appuyant sur l’analyse des mouvements linéaires du corps. Cependant, le maintien de l’équilibre au cours de nos différentes tâches motrices nécessite à la fois un contrôle approprié des mouvements linéaires et des mouvements angulaires (de rotation) des segments corporels. Ainsi, ce travail doctoral vise à identifier et comprendre les modifications du contrôle des mouvements angulaires du corps chez les personnes âgées à travers l’étude du moment cinétique du corps, qui est reconnu comme étant un paramètre mécanique hautement contrôlé par le système nerveux central pour le maintien de l’équilibre. Nos études mettent en évidence qu’au cours de la tâche d’exécution du pas volontaire, les personnes âgées présentent une modification dans le contrôle du moment cinétique du corps. Globalement, nos résultats montrent que les personnes âgées ont une amplitude du moment cinétique du corps plus importante que des individus jeunes - en particulier dans le plan sagittal -, et que cela est exacerbé avec l’augmentation de la vitesse d’exécution du mouvement (ou de progression). En outre, nos résultats mettent en exergue que cette modification dans le contrôle du moment cinétique du corps avec l’âge est directement imputée aux changements au niveau des moments cinétiques segmentaires, avec les personnes âgées présentant des moments cinétiques du tronc et des membres inférieurs plus élevés que les adultes jeunes. Pour finir, nous avons observé une relation entre l’amplitude du moment cinétique du corps au cours de la tâche d’exécution du pas volontaire et des mesures de force musculaire et d’équilibre chez des participants jeunes et âgés. Ces résultats suggèrent que les modifications dans le contrôle du moment cinétique du corps avec l’âge, qui pourraient imposer une plus grande difficulté pour le contrôle de l’équilibre et potentiellement un plus grand risque de chute lors de la tâche d’exécution du pas volontaire chez la personne âgée, peuvent être en partie attribuées à la perte de force musculaire des membres inférieurs et à la détérioration des autres systèmes impliqués dans le contrôle de l’équilibre avec le vieillissement. Ensemble, nos travaux concourent à une meilleure compréhension des mécanismes sous-jacents l’altération du contrôle de l’équilibre chez la personne âgée et offrent une base à de futures études visant à réduire l’incidence des chutes au sein de cette même population
In order to understand the etiology of falls in the elderly, studies in the field of biomechanics have mainly focused on assessing balance abilities, relying on the analysis of linear body movements. However, maintaining balance during our various motor tasks requires both appropriate control of linear movements and angular (rotational) movements of body segments. Thus, this doctoral work aims to identify and understand changes in the control of angular body movements in old adults through the study of whole-body angular momentum, which is recognized as a mechanical parameter highly controlled by the central nervous system to maintain balance. Our studies reveal that during the volitional stepping task, old adults exhibit an alteration in the control of the whole-body angular momentum. Overall, our results show that old adults have a higher range of whole-body angular momentum than their younger counterparts – particularly in the sagittal plane – and this is exacerbated with the increase in progression velocity. Furthermore, our results highlight that this age-related changes in the control of whole-body angular momentum is directly ascribed to changes in segmental angular momenta, with old adults having higher trunk and lower limb angular momenta than young adults. Finally, we observed a relationship between the range of whole-body angular momentum during the stepping task and measurements of muscle strength and balance in young and old adults. These results suggest that age-related changes in the control of whole-body angular momentum, which may impose greater challenge for balance control and potentially a greater risk of fallingduring the volitional stepping task in old adults, can be partially attributed to the decline in lower limb muscle strength and the impairment of other systems involved in balance control with aging. Together, our results contribute to a better understanding of the mechanisms underlying impaired balance control in old adults and provide a basis for future studies to reduce the incidence of falls in this population

Dans le cadre de la régression, de nombreuses études s’intéressent au problème dit de la grande dimension, où le nombre de variables explicatives mesurées sur chaque échantillon est beaucoup plus grand que le nombre d’échantillons. Si la sélection de variables est une question classique, les méthodes usuelles ne s’appliquent pas dans le cadre de la grande dimension. Ainsi, dans ce manuscrit, nous présentons la transposition de tests statistiques classiques à la grande dimension. Ces tests sont construits sur des estimateurs des coefficients de régression produits par des approches de régressions linéaires pénalisées, applicables dans le cadre de la grande dimension. L’objectif principal des tests que nous proposons consiste à contrôler le taux de fausses découvertes. La première contribution de ce manuscrit répond à un problème de quantification de l’incertitude sur les coefficients de régression réalisée sur la base de la régression Ridge, qui pénalise les coefficients de régression par leur norme l2, dans le cadre de la grande dimension. Nous y proposons un test statistique basé sur le rééchantillonage. La seconde contribution porte sur une approche de sélection en deux étapes : une première étape de criblage des variables, basée sur la régression parcimonieuse Lasso précède l’étape de sélection proprement dite, où la pertinence des variables pré-sélectionnées est testée. Les tests sont construits sur l’estimateur de la régression Ridge adaptive, dont la pénalité est construite à partir des coefficients de régression du Lasso. Une dernière contribution consiste à transposer cette approche à la sélection de groupes de variables
In the regression framework, many studies are focused on the high-dimensional problem where the number of measured explanatory variables is very large compared to the sample size. If variable selection is a classical question, usual methods are not applicable in the high-dimensional case. So, in this manuscript, we develop the transposition of statistical tests to the high dimension. These tests operate on estimates of regression coefficients obtained by penalized linear regression, which is applicable in high-dimension. The main objective of these tests is the false discovery control. The first contribution of this manuscript provides a quantification of the uncertainty for regression coefficients estimated by ridge regression in high dimension. The Ridge regression penalizes the coefficients on their l2 norm. To do this, we devise a statistical test based on permutations. The second contribution is based on a two-step selection approach. A first step is dedicated to the screening of variables, based on parsimonious regression Lasso. The second step consists in cleaning the resulting set by testing the relevance of pre-selected variables. These tests are made on adaptive-ridge estimates, where the penalty is constructed on Lasso estimates learned during the screening step. A last contribution consists to the transposition of this approach to group-variables selection

La tesi analizza il contemporaneo declino elettorale del Partito Laburista israeliano in relazione agli eventi storici, ai cambiamenti sociali e demografici che hanno portato ad un effettivo cambiamento del sistema politico. In particolar modo la ricerca si sofferma sulla lettura dei risultati elettorali, cercando di sottolineare come le dinamiche sopra indicate abbiano influenzato i trend elettorali e l'offerta politica del partito stesso. Fino a giungere agli anni Novanta, passaggio fondamentale in cui cogliere le ragioni per le quali il Partito Laburista sembra tutt'ora non riuscire invertire il trend negativo degli ultimi quindici anni.
This thesis analyzes the contemporary electoral decline of the Israeli Labor Party in relation to historical events, social and demographic changes that have led to an effective change in the Israeli political system. In particular, the research focuses on the electoral results, trying to underline how the dynamics indicated above have influenced the electoral trends and the political offer of the party itself. The Nineties represent a fundamental passage in which it is possibleto understand the reasons why the Labor Party seems unable, still today, to reverse the negative trend of the last fifteen years.

Еолски седименти представљају идеалан материјал за примену методе луминесцентног датирања. Један од главних разлога је њихов минеролошки састав који углавном чине кварц и фелдспат. Ови минерали поседују читав низ луминесцентних карактеристика који их чине погодним за луминесцентно датирање а поред тога, могуће је релативно једноставно извршити њихову екстракцију и сепарацију од других неминералних састојака. У случају лесних платоа (еолски транспортованих седимената), може се претпоставити да су индивидуална минерална зрна била потпуно изложена сунчевој светлости и да су у довољној мери ресетована пре депозиције. Истраживана секција Велики сурдук на Тителском лесном платоу, изабрана је за датирање, на основу своје моћности, високе стопе седиментације и комплетности које овај локалитет чине најдетаљнијим палеоклиматским архивом последња два глацијално- интерглацијална циклуса у овом делу Европе.Употребом класичног SAR протокола за датирање утврђено је да кварцна зрна саТителског лесног платоа представљају прецизне дозиметре до ~120 Gy, после чега долази до очигледног засићења сигнала. Највиша утврђена доза фелдспата измерена модификованим SAR пост-IRIR^200,290 протоколом је била 854,0±24 Gy. Пост-IRIR^200,290 протокол није утврдио сатурацију код минералних зрна фелдспата, што је умногомо подигло поуздану старосну границу за датирање на секцији Велики сурдук, а највероватније и на осталим лесним профилима у Србији. Примена различитих протокола и мерења аплицираних на различите групе минерала и њихове фракције дају овим истраживањима далекосежан методолошки значај. На овај начин је доказано да у поменутом временском опсегу OSL датирање грубог кварца даје поуздане старосне процене до 35,8±3,7 ka док су зрна фелдспата датирана до чак 237,7±12,8 ka. Кварц има апсолутно највећу подударност са очекиваним геолошким временом до MIS 2 за разлику од фелдспата који показује бoљу подударност са претпостављеном геолошком хронологијом до MIS 7.
Eolski sedimenti predstavljaju idealan materijal za primenu metode luminescentnog datiranja. Jedan od glavnih razloga je njihov minerološki sastav koji uglavnom čine kvarc i feldspat. Ovi minerali poseduju čitav niz luminescentnih karakteristika koji ih čine pogodnim za luminescentno datiranje a pored toga, moguće je relativno jednostavno izvršiti njihovu ekstrakciju i separaciju od drugih nemineralnih sastojaka. U slučaju lesnih platoa (eolski transportovanih sedimenata), može se pretpostaviti da su individualna mineralna zrna bila potpuno izložena sunčevoj svetlosti i da su u dovoljnoj meri resetovana pre depozicije. Istraživana sekcija Veliki surduk na Titelskom lesnom platou, izabrana je za datiranje, na osnovu svoje moćnosti, visoke stope sedimentacije i kompletnosti koje ovaj lokalitet čine najdetaljnijim paleoklimatskim arhivom poslednja dva glacijalno- interglacijalna ciklusa u ovom delu Evrope.Upotrebom klasičnog SAR protokola za datiranje utvrđeno je da kvarcna zrna saTitelskog lesnog platoa predstavljaju precizne dozimetre do ~120 Gy, posle čega dolazi do očiglednog zasićenja signala. Najviša utvrđena doza feldspata izmerena modifikovanim SAR post-IRIR^200,290 protokolom je bila 854,0±24 Gy. Post-IRIR^200,290 protokol nije utvrdio saturaciju kod mineralnih zrna feldspata, što je umnogomo podiglo pouzdanu starosnu granicu za datiranje na sekciji Veliki surduk, a najverovatnije i na ostalim lesnim profilima u Srbiji. Primena različitih protokola i merenja apliciranih na različite grupe minerala i njihove frakcije daju ovim istraživanjima dalekosežan metodološki značaj. Na ovaj način je dokazano da u pomenutom vremenskom opsegu OSL datiranje grubog kvarca daje pouzdane starosne procene do 35,8±3,7 ka dok su zrna feldspata datirana do čak 237,7±12,8 ka. Kvarc ima apsolutno najveću podudarnost sa očekivanim geološkim vremenom do MIS 2 za razliku od feldspata koji pokazuje bolju podudarnost sa pretpostavljenom geološkom hronologijom do MIS 7.
Aeolian sediments represent an ideal material for the application of the luminescence dating method. One of the main reasons is their mineralogical composition which consists mainly of quartz and feldspars. These minerals possess a whole range of luminescence characteristics that make them suitable for luminescence dating, and in   addition, it is relatively easy to extract and separate them from other non-mineral ingredients. In the case of loess plateaus (eolically transported sediments), it can be assumed that individual mineral grains were completely exposed to sunlight and were sufficiently bleached before deposition. The investigated section Veliki surduk on the Titel loess plateau was chosen for dating, based on its thicknes, the high rate of sedimentation and completeness which makes this site one of the most detailed paleoclimatic archives of the last two glacial and interglacial cycles in this part of Europe.Using the classic SAR protocol for dating, it has been found that the quartz grains from the Titel loess plateau are precise dosimeters up to ~120 Gy, followed by an apparent saturation of the signal. The highest established dose of feldspars measured by the modified SAR postIRIR^200,290 protocol was 854.0±24 Gy. The post-IRIR^200,290 protocol did not determine the saturation of feldspar minerals, which has largely raised the age limit for luminescence dating at the Veliki surduk section, and most probably, also on other loess profiles in Serbia. The application of various protocols and measurements applied to different groups of minerals and their fractions give these research a far-reaching methodological significance. In this way, it has been proven that in the mentioned time range, OSL dating of coarse grain quartz yields reliable age estimates up to 35.8±3.7 k while the feldspar grains are dated to as high as 237.7±12.8 ka. Quartz has the highest match with the expected geological age to MIS 2 as opposed to feldspar, which shows better correspondance with the assumed geological chronology to MIS 7.

Cette thèse vise à élaborer de nouvelles stratégies de commande basées sur la commande prédictive pour le système de génération d’énergie éolienne. La topologie des systèmes de production éolienne basées sur le Générateur Asynchrone à Double Alimentation (GADA) qui convient à des plateformes de génération dans la gamme de puissance de 1.5 à 6 MW est abordée. Du point de vue technologique, le convertisseur à trois niveaux et clampé par le neutre (3L-NPC) est considéré comme une bonne solution pour une puissance élevée en raison de ses avantages: capacité à réduire la distorsion harmonique de la tension de sortie et du courant, et augmentation de la capacité du convertisseur grâce à une tension réduite appliquée à chaque semi-conducteur de puissance. Une description détaillée de la commande prédictive à ensemble de commande fini (FCS-MPC) avec un horizon de prédiction de deux pas est présentée pour deux boucles de régulation: celle liée au convertisseur connecté au réseau et celle du convertisseur connecté au GADA. Le principe de la commande repose sur l’utilisation d’un modèle de prédiction permettant de prédire le comportement du système pour chaque état de commutation du convertisseur. La minimisation d’une fonction de coût appropriée prédéfinie permet d’obtenir la commutation optimale à appliquer au convertisseur. La thèse étudie premièrement les problèmes liées à la compensation du temps de calcul de la commande et au choix et aux pondérations de la fonction de coût. Ensuite, le problème de stabilité de la commande FCS-MPC est abordé en considérant une fonction de Lyapunov dans la minimisation de la fonction de coût. Finalement, une étude sur la compensation des effets des temps morts du convertisseur est présentée
This thesis aims to elaborate new control strategies based on Model Predictive control for wind energy generation system. We addressed the topology of doubly fed induction generator (DFIG) based wind generation systems which is suitable for generation platform power in the range in 1.5-6 MW. Furthermore, from the technological point of view, the three-level neutral-point clamped (3L-NPC) inverter configuration is considered a good solution for high power due to its advantages: capability to reduce the harmonic distortion of the output voltage and current, and increase the capacity of the converter thanks to a decreased voltage applied to each power semiconductor.In this thesis, we presented a detailed description of finite control set model predictive control (FCS-MPC) with two step horizon for two control schemes: grid and DFIG connected 3L-NPC inverter. The principle of the proposed control scheme is to use system model to predict the behaviour of the system for every switching states of the inverter. Then, the optimal switching state that minimizes an appropriate predefined cost function is selected and applied directly to the inverter.The study of issues such as delay compensation, computational burden and selection of weighting factor are also addressed in this thesis. In addition, the stability problem of FCS-MPC is solved by considering the control Lyapunov function in the design procedure. The latter study is focused on the compensation of dead-time effect of power converter

Dans cette thèse, nous examinons plusieurs aspects de l'estimation des paramètres pour les statistiques et les techniques d'apprentissage automatique, aussi que les méthodes d'optimisation applicables à ces problèmes. Le but de l'estimation des paramètres est de trouver les paramètres cachés inconnus qui régissent les données, par exemple les paramètres dont la densité de probabilité est inconnue. La construction d'estimateurs par le biais de problèmes d'optimisation n'est qu'une partie du problème, trouver la valeur optimale du paramètre est souvent un problème d'optimisation qui doit être résolu, en utilisant diverses techniques. Ces problèmes d'optimisation sont souvent convexes pour une large classe de problèmes, et nous pouvons exploiter leur structure pour obtenir des taux de convergence rapides. La première contribution principale de la thèse est de développer des techniques d'appariement de moments pour des problèmes de régression non linéaire multi-index. Nous considérons le problème classique de régression non linéaire, qui est irréalisable dans des dimensions élevées en raison de la malédiction de la dimensionnalité. Nous combinons deux techniques existantes : ADE et SIR pour développer la méthode hybride sans certain des aspects faibles de ses parents. Dans la deuxième contribution principale, nous utilisons un type particulier de calcul de la moyenne pour la descente stochastique du gradient. Nous considérons les familles exponentielles conditionnelles (comme la régression logistique), où l'objectif est de trouver la valeur inconnue du paramètre. Nous proposons le calcul de la moyenne des paramètres de moments, que nous appelons fonctions de prédiction. Pour les modèles à dimensions finies, ce type de calcul de la moyenne peut entraîner une erreur négative, c'est-à-dire que cette approche nous fournit un estimateur meilleur que tout estimateur linéaire ne peut jamais le faire. La troisième contribution principale de cette thèse porte sur les pertes de Fenchel-Young. Nous considérons des classificateurs linéaires multi-classes avec les pertes d'un certain type, de sorte que leur double conjugué a un produit direct de simplices comme support. La formulation convexe-concave à point-selle correspondante a une forme spéciale avec un terme de matrice bilinéaire et les approches classiques souffrent de la multiplication des matrices qui prend beaucoup de temps. Nous montrons que pour les pertes SVM multi-classes avec des techniques d'échantillonnage efficaces, notre approche a une complexité d'itération sous-linéaire, c'est-à-dire que nous devons payer seulement trois fois O(n+d+k) : pour le nombre de classes k, le nombre de caractéristiques d et le nombre d'échantillons n, alors que toutes les techniques existantes sont plus complexes
In this thesis we consider several aspects of parameter estimation for statistics and machine learning and optimization techniques applicable to these problems. The goal of parameter estimation is to find the unknown hidden parameters, which govern the data, for example parameters of an unknown probability density. The construction of estimators through optimization problems is only one side of the coin, finding the optimal value of the parameter often is an optimization problem that needs to be solved, using various optimization techniques. Hopefully these optimization problems are convex for a wide class of problems, and we can exploit their structure to get fast convergence rates. The first main contribution of the thesis is to develop moment-matching techniques for multi-index non-linear regression problems. We consider the classical non-linear regression problem, which is unfeasible in high dimensions due to the curse of dimensionality. We combine two existing techniques: ADE and SIR to develop the hybrid method without some of the weak sides of its parents. In the second main contribution we use a special type of averaging for stochastic gradient descent. We consider conditional exponential families (such as logistic regression), where the goal is to find the unknown value of the parameter. Classical approaches, such as SGD with constant step-size are known to converge only to some neighborhood of the optimal value of the parameter, even with averaging. We propose the averaging of moment parameters, which we call prediction functions. For finite-dimensional models this type of averaging can lead to negative error, i.e., this approach provides us with the estimator better than any linear estimator can ever achieve. The third main contribution of this thesis deals with Fenchel-Young losses. We consider multi-class linear classifiers with the losses of a certain type, such that their dual conjugate has a direct product of simplices as a support. We show, that for multi-class SVM losses with smart matrix-multiplication sampling techniques, our approach has an iteration complexity which is sublinear, i.e., we need to pay only trice O(n+d+k): for number of classes k, number of features d and number of samples n, whereas all existing techniques have higher complexity

Uvod: Infertilitet pogađa 10-15% parova reproduktivnog doba. Vanetesna oplodnja (VTO) je najefikasniji vid tret-mana infertiliteta, ali uprkos značajnom napretku stopa uspeha VTO u proseku iznosi oko 30% po ciklusu. Glavnim razlogom neuspeha smatra se neadekvatan kvalitet embriona, dok se pretpostavlja da u 10-20% slučajeva razlog neuspeha leži u neadekvatnoj receptivnosti uterusa. Na osnovu inicijalnih istraživanja histeroskopija, koja predstvalja zlatni standard u dijagnostici i tretmanu patologije kavuma uterusa, se često izvodi u svakodnevnoj kliničkoj praksi kako bi se povećala uspešnost VTO. Uprkos širokoj primeni i dalje ne postoji dovoljno kvalitetnih dokaza o realnoj ulozi histeroskopije na ishod VTO kako kod patoloških stanja kavuma tako i rutinski, pre prvog ili rekurentnog pokušaja VTO. Cilj disertacije bio je da se utvrdi uticaj sprovođenja histeroskopije na ishod VTO, ustanovi učestalost prethodno neprepoznate patologije kavuma uterusa, kao i da se ispitaju stavovi pacijenata o primeni rutinske histeroskopije pred VTO. Materijal i metode: Istraživanje je sprovedeno u Kliničkom centru Vojvodine, u formi prospektivne studije u dve sukcesivne etape od 01.01.2015. do 01.04.2017. U prvoj etapi poređen je ishod VTO kod pacijentkinja kojima pred postupak VTO nije sprovedena histeroskopija (Grupa A), pacijentkinja kod kojih je dobijen uredan nalaz histeroskopije pred postupak VTO (Grupa B) i pacijentkinja gde je pred postupak VTO dobijen patološki nalaz kavuma na histeroskopiji koji je u istom aktu tertian (Grupa C). Druga etapa istraživanja predstavljala je randomiziranu kontrolisanu studiju (RCT – randomised controlled trial). Nakon verifikacije urednog ultrazvučnog nalaza pred prvi postupak VTO, pacijentkinje su randomizirane u Grupu A2 kojima pred postupak VTO nije sprovedena histeroskopija i Grupu B2 kojima je pred postupak VTO sprovedena rutinska histeroskopija. Statistička analiza sprovedena je upotrebom odgovarajućeg softvera (JMP Ver. 9). Poređeni su podaci o osnovnim karakteristikama pacijenata, toka i ishoda ciklusa VTO. Primarni parametar ishoda bila je stopa kliničke trudnoće po embriotransferu. Pored analize ishoda primarno konstruisanih grupa, urađena je analiza i naknadno konstruisanih subgrupa, kao i predikcioni model uspeha VTO baziran na logističkoj regresiji. Rezultati: Studija je uključila 253 pacijentkinje (52 pacijentkinja iz Grupe A, 50 iz Grupe B, 50 iz Grupe C, 51 iz Grupe A2 i 50 iz Grupe B2). Nije postojala statistički značajna razlika u karakteristikama pacijentkinja, parametrima ovarijalne rezerve, broju dobijenih jajnih ćelija ni drugim parametrima toka postupka VTO među posmatranim grupama. U prvoj etapi istraživanja dobijena je statistički značajno (p=0,013) veća stopa kliničkih trudnoća kod pacijentkinja kojima je pred postupak VTO sprovedena histeroskopija - 50 % za Grupu B i 42% za grupu C u odnosu na 30,77% kod pacijentkinja bez histeroskopije (Grupa A), bez statistički značajne razlike među histeroskopskim grupama. U drugoj etapi istraživanja stopa kliničkih trudnoća prilikom upotrebe rutinske histeroskopije pred prvu VTO (Grupa B2) iznosila je 46% naspram 31,37% kod pacijentkinja bez histeroskopije pred prvu VTO (Grupa A2), iako uočena razlika nije dostigla statističku značajnost (p =0,089), uz relativan rizik (RR) za ostvarivanje kliničke trudnoće nakon primene histeoskopije uiznosio od 1,47 (95% CI 0,88-2,43) (p=0,13). Analizon subgrupa kod 100 pacijentkinja sa rutinski sprovedenom histeroskopijom pred VTO i 103 pacijentkinje bez histeroskopije pred VTO, dobijena je statistički značajnao veća stopa kliničkih trudnoća (48% naspram 31,07%, istim redom), uz RR od 1,54 (95% CI 1,08-2,20) (p=0,013), kao i stopa tekućih trudnoća od RR 1,49 (CI 1,01-2,19) (p= 0,039). Analiza ukupnog uticaja izvođenja histeroskopije pred VTO dobila je statistički značanjno veću stopu kliničkih trudnoća po ET za grupu histeroskopije uz RR 1,48 (CI 1,06-2,07) (p=0,017). Histeroskopijom je nakon urednog ultrazvučnog nalaza ustanovljeno postojanje patološkog nalaza kod 34,65% pacijenata i to 22,7% major patologije i 11,88% minor patologije kavuma. Nije postojala statistički značajna razlika u uspehu VTO u odnosu na sam nalaz histeroskopije. 98,67% pacijenata podržalo je rutinsku upotrebu histeroskopije pred prvi postupak VTO, dok je 83% pacijenata podržavlo rutinsku upotrebu histeroskopije pred svaki postupak VTO. U finalnom predikcionom modelu se uz AUC od 0,748 jedino postojanje visokokvalitetnog embriona uz odnos šansi (OR) 7,91 (95% CI 1,80-56,06; p=0,0047), transfer blastociste uz OR 3,80 (95% CI 1,90-7,98; p=0,0001) i izvođenje histeroskopije pred VTO uz OR 2,13 (95% CI 1,14-4,08, p=0,0169) pokazalo statistički značajnim prediktorima trudnoće. Diskusija: Studija je dobila pozitivan uticaj histeroskopije na ishod postupka VTO, iskazan pre svega povećanjem stope kliničkih trudnoća nakon sprovođenja histeroskopije (bilo da je na histeroskopiji nađen uredan ili patološki nalaz). Dodatna prednost histeroskopije predstavljala je i i detekcija prethodno nepropoznate patologije kavuma. Umeren efekat na ukupno poboljšanje stope kliničkih trudnoća prilikom rutinskog sprovođenja histeroskopije pred prvu VTO, koji je statističku značajnost dostigao tek analizom subgrupa u skladu je sa nalazima novijih dobro dizajniranih studija koji donekle limitiraju nekritičku upotrebu histeroskopije. Biološko objašnjenje potencijalnog pozitivnog uticaja histeroskopije najverovatnije leži u detekciji i tretmanu prethodno nepropoznate patologije kavuma, olakšavanju procedure embriotransfera, kao i humoralnim i molekularnim promenama koje nastaju u endometrijumu kao posledica odgovarajuće histeroskopske traume a koji su u dosašanjim istraživanjima apostrofirani kao faktori koji mogu povećati receptivnost uterusa. Zaključak: Histeroskopija je efikasna, bezbedna i visoko prihvatljiva procedura koja dovodi do povećanja uspeha VTO u standardnim kliničkim indikacijama (prethodnog neuspelog postupka VTO i sumnje na patološki nalaz kavuma uterusa) bilo da se na samoj histeroskopiji nađe uredan ili patološki nalaz. Rutinska primena histeroskopije pred prvi postupak VTO se na osnovu rezultata studije ne može smatrati apsolutno opravdanom usled statistički nedovoljno značajnog povećanja stope kliničke trudnoće. Uzevši u obzir visoku prihvatljivost od strane pacijenata i najverovatniji pozitivan efekat na stopu trudnoće primena rutinske histeroskopije pred prvu VTO bila bi opravdana ukoliko se implementira koncept ambulantne histeroskopije.
Introduction: Infertility affects 10-15% of all couples. In vitro fertilisation (IVF) is the most effective method of infertility treatment, but despite a significant improvement, success rate of IVF is still around 30% per cycle. The main reason for the IVF failure is inadequate embryo quality, but in 10-20% of cases the cause of IVF failure lies in impaired uterine receptivity. Based on earlier studies hysteroscopy, gold standard in the diagnosis and treatment of uterine cavity pathology, is often performed to increase IVF success. Despite its wide use, there is lack of high quality evidence regarding real contribution of hysteroscopy on IVF outcome in situations of uterine cavity pathology or routinely prior to first IVF or after recurrent implantation failure. The aim of this dissertation was to determine the influence of performing hysteroscopy on IVF outcome, as well as the incidence of previously unrecognized uterine pathology, and to examine patient's attitudes about performing routine hysteroscopy prior to IVF. Material and methods: The research was conducted in a prospective manner in two successive stages at Clinical Center of Vojvodina from 01.01.2015. until 01.04.2017. During first stage of the study IVF outcome was compared between patients who did not have a hysteroscopy prior to IVF (group A), patients with normal hysteroscopic finding prior to the IVF (Group B) and patients with abnormal hysteroscopic findings prior to IVF which was treated at the same time (Group C). The second stage of the study was a randomized controlled trial (RCT). After verification of normal ultrasound findings prior to the first IVF, patients were randomized to group A2 in who me hysteroscopy was not performed and group B2 who had routine hysteroscopy prior to first IVF. Statistical analysis was carried out using the appropriate statistical software (JMP Ver. 9). Patient characteristics, course and outcome of IVF cycle were compared between groups. The primary outcome was clinical pregnancy rate (CPR) per embryotransfer. In addition to analyzing the IVF outcomes in primarily defined groups, subgroup analysis was also performed, as well as IVF success pre-diction model based on logistic regression. Results: The study included 253 patients (52 patients in Group A, 50 in Group B, 50 in Group C, 51 in Group A2 and 50 in Group B2). There was no statistically significant difference in patient characteristics, ovarian reserve parameters, number of retrieved oocytes, or other relevant parameters of IVF course between the observed groups. In the first stage of the study there was statistically significant (p = 0.013) higher clinical pregnancy rate in patients who had a hysteroscopy before IVF - 50% for Group B and 42% for group C versus 30,77 % in patients without hysteroscopy before IVF (Group A), without statistically significant difference between hysteroscopic groups. In the second stage of the study, routine hysteroscopy prior to first IVF (Group B2) led to clinical pregnancy rate 46% versus 31.37% in patients without hysteroscopy prior to first IVF (Group A2), although without statistical significance (p = 0.089. Relative risk (RR) for achieving clinical pregnancy after performing hysteroscopy was 1.47 (95% CI 0.88-2.43) (p = 0.13). Subgroup analysis of 100 patients with routinely performed hysteroscopy before IVF and 103 patients without hysteroscopy prior to the IVF showed statistically significant higher rates of clinical pregnancies (48% versus 31.07%, in the same order), with RR of 1.54 (95% CI 1.08-2.20), (p = 0.013), and for ongoing pregnancies RR was 1.49 (95% CI 1.01-2.19) (p = 0.039). Overall effect of performing hysteroscopy prior to IVF resulted in a statistically significant increase in the clinical pregnancy with RR 1.48 (95% CI 1.06-2.07) (p = 0.017). After normal ultrasound finding hysteroscopy revealed 34.65% of pathological finding, 22.7% of major and 11.88% of minor pathology of the cavity). There was no statistically significant difference in IVF outcome based on hysteroscopy findings. 98.67% of patients supported the routine use of hysteroscopy before the first IVF procedure, while 83% of patients supported the routine use of the hysteroscopy before every IVF procedure. In the final prediction model, with the AUC of 0.748, only the presence of high quality embryos with odds ratio (OR) 7,91 (95% CI 1,80-56,06; p=0,0047), blastocyst transfer with OR 3,80 (95% CI 1,90-7,98; p=0,0001) and performing hysteroscopy prior to IVF with OR 2,13 (95% CI 1,14-4,08, p=0,0169) proved to be statistically significant predictors of pregnancy. Discussion: The study shoved a positive influence of hysteroscopy on the IVF outcome by increasing clinical pregnancy rate after performing hysteroscopy (whether hysteroscopy revealed normal or pathological finding). Additional benefit of hysteroscopy was detection of previously unrecognized uterine pathology. A moderate effect on the overall improvement in clinical pregnancy rate with use of routine hysteroscopy, which reached statistical significance only by subgroup analysis, is in line with findings of recent well designed studies that somewhat limit the noncritical use of hysteroscopy. A biological explanation of the potential positive effect of hysteroscopy is most likely due to detection and treatment of the previously unrecognized uterine pathology, facilitating embryotransfer procedure, as well as the humoral and molecular changes that occur in the endometrium as a consequence of the hysteroscopic trauma. Those changes were hypothesized as factors that can increase uterine receptivity by numerous research. Conclusion: Hysteroscopy is an effective, safe and highly acceptable procedure that increases IVF success when performed for accepted clinical indications (previous IVF failures, pathological findings of uterine cavity), whether hysteroscopy reveals normal or pathological finding. The routine use of hysteroscopy prior to first IVF based on this study can not be considered justified since increase in clinical pregnancy rate did not reach statistical significance. Given the high acceptance of this concept by the patients and moderate but probable positive effect on IVF outcome, implementation of routine hysteroscopy prior to first VTO would be justified only in office hysteroscopy setting.

Diploma thesis solves the creating of layout study and creation of documentation for the realization new fire station. In the object is contemplated with placement of regional fire brigade unit, category JPO I and volunteer firefighter unit of city together, category JPO III/1. Fire station category will be considered as P1. The object consists of two structurally and functionally different parts. The first part is made of ceramic masonry with external thermal insulation composite system. Here are located administrative and operational premises. This section has two floors. Second part is solved as single-storey skeletal system made of reinforced concrete, roofed by prestressed concrete trusses. This part contains technical facilities and a garage. The cladding of the skeletal part of building is designed by using sandwich facade panels. Object has dissected ground plan, maximum dimensions are 54,80x22,20 m. Building has no basement and is roofed by flat roofs only. Object is located in peripheral locality of the city Jaroměř. Surrounding terrain is flat. Solved area contains a few related objects.

Ultra Wideband (UWB) communication has recently drawn considerable attention from academia and industry. This is mainly owing to the ultra high speeds and cognitive features it could offer. The employability of UWB in numerous areas including but not limited to Wireless Personal Area Networks, WPAN's, Body Area Networks, BAN's, radar and medical imaging etc. has opened several avenues of research and development. However, still there is a disagreement on the standardization of UWB. Two contesting radios for UWB are Multiband Orthogonal Frequency Division Multiplexing (MB-OFDM) and DS-UWB (Direct Sequence Ultra Wideband). As nearly all of the reported research on UWB hasbeen about a very narrow/specific area of the communication system, this thesis looks at the end-to-end performance of an MB-OFDM approach. The overall aim of this project has been to first focus on three different aspects i.e. interference, antenna and propagation aspects of an MB-OFDM system individually and then present a holistic or an end-to-end system analysis finally. In the first phase of the project the author investigated the performance of MB-OFDM system under the effect of his proposed generic or technology non-specific interference. Avoiding the conventional Gaussian approximation, the author has employed an advanced stochastic method. A total of two approaches have been presented in this phase of the project. The first approach is an indirect one which involves the Moment Generating Functions (MGF's) of the Signal-to-Interference-plus-Noise-Ratio (SINR) and the Probability Density Function (pdf) of the SINR to calculate the Average Probabilities of Error of an MB-OFDM system under the influence of proposed generic interference. This approach assumed a specific two-dimensional Poisson spatial/geometric placement of interferers around the victim MB-OFDM receiver. The second approach is a direct approach and extends the first approach by employing a wider class of generic interference. In the second phase of the work the author designed, simulated, prototyped and tested novel compact monopole planar antennas for UWB application. In this phase of the research, compact antennas for the UWB application are presented. These designs employ low-loss Rogers duroid substrates and are fed by Copla-nar Waveguides. The antennas have a proposed feed-line to the main radiating element transition region. This transition region is formed by a special step-generating function-set called the "Inverse Parabolic Step Sequence" or IPSS. These IPSS-based antennas are simulated, prototyped and then tested in the ane-choic chamber. An empirical approach, aimed to further miniaturize IPSS-based antennas, was also derived in this phase of the project. The empirical approach has been applied to derive the design of a further miniaturized antenna. More-over, an electrical miniaturization limit has been concluded for the IPSS-based antennas. The third phase of the project has investigated the effect of the indoor furnishing on the distribution of the elevation Angle-of-Arrival (AOA) of the rays at the receiver. Previously, constant distributions for the AOA of the rays in the elevation direction had been reported. This phase of the research has proposed that the AOA distribution is not fixed. It is established by the author that the indoor elevation AOA distributions depend on the discrete levels of furnishing. A joint time-angle-furnishing channel model is presented in this research phase. In addition, this phase of the thesis proposes two vectorial or any direction AOA distributions for the UWB indoor environments. Finally, the last phase of this thesis is presented. As stated earlier, the overall aim of the project has been to look at three individual aspects of an MB-OFDM system, initially, and then look at the holistic system, finally. Therefore, this final phase of the research presents an end-to-end MB-OFDM system analysis. The interference analysis of the first phase of the project is revisited to re-calculate the probability of bit error with realistic/measured path loss exponents which have been reported in the existing literature. In this method, Gaussian Quadrature Rule based approximations are computed for the average probability of bit error. Last but not the least, an end-to-end or comprehensive system equation/impulse response is presented. The proposed system equation covers more aspects of an indoor UWB system than reported in the existing literature.

This diploma thesis solves creating of the layout study and creation of documentation for the realization of a new fire station. In the building, the location of the JPO I rescue units is considered. The territorial jurisdiction is usually considered to be within 20 minutes of dislocation. The category of the fire station will be considered as P1. The object consists of two structurally and functionally different parts. The first part is designed as a one-storey reinforced concrete skeleton roofed with pre-stressed ceiling panels. This section includes technical background and garages. The cladding of the skeletal part of the building is solved by a masonry block made of lime-sand blocks with external thermal insulation composite system. The second part is walled from lime-sand blocks with external thermal insulation composite system. There are spaces of administrative and operational character. This section has two above-ground floors. The ground plan is fractured, the maximum dimensions are 45.59 x 28.34 m. The building has no basement, roofed only with flat roofs. The object is situated in the peripheral locality of Příbor, the terrain is slightly sloping.

碩士
東吳大學
經濟學系
100
Investors based on their investment objectives and risk tolerance of the basic understanding of the establishment of the stop profit and stop-loss point mechanism to protect their investment results. In when do financial investment, often will set a stop-loss point, stop profit points to set a stop-loss points to limit the magnitude of the losses within the acceptable range, and when you reach a certain income, will set up a stop profit points to avoid the market overheating or the fierce market shocks not only did not make money, also stuck. When investors face uncertain risk in the stock market, how to determine its own optimum proportion of funds, and consider the stop-loss ahead of appearances profit-taking and stop profit set to pursue a great rate of return and risk control. The Kelly investment approach in the case of control risk or no control risk, investment in the Taiwan stock index and its performance are backward buy-and-hold strategy, Kelly investment approach that technical analysis methods, the use of reward probability assignedas a simulation of future probability distribution, less suitable for use in the Taiwan Weighted Stock Index above, because TAIEX after multiple cyclical fluctuations. The optimal rate of investment in several major parameters assumed to seek out fixed value, but also affect the geometric average rate of return of the investment. Under different investment ratios, these parameter values are not necessarily the most appropriate, by adding the genetic algorithm is applied to solve the optimization parameters, you should be able to improve its rate of return performance.

The aim of the thesis was to optimize selected trading strategy in the currency market FOREX. For the optimization was chosen strategy Donchian 5 and 20 and was applied on currency pair EUR/USD. At first was performed backtesting, exit optimization, time optimization and found the importance of position sizing. Then the system was tested in paper trading at broker Admiral Markets, daytrading five minutes chart. Business strategy was optimized and is achieving positive results.

碩士
東吳大學
資訊管理學系
101
There are three trading strategies in this study. The first one is a buy and hold strategy (Buy & Hold) used as a reference value comparison. The reversal patterns method, which adopts 6 kinds of candlestick form as trading signals, is the second. Seven consecutive 5-day average lines are used to judge for the tendency. We buy long when reverse signals rise in the downward trend and sell short when reverse signals appear in an upward trend. In the buying and selling strategy, the same signals are ignored after we have performed his action. Its performance is compared with the first method. The third method is based on the stop-loss or stop-profit to improve the reversal patterns method. In our experiments, the effect of the third method is quite good. Furthermore the performance of the method relying only stop-profit point greatly exceeds the one of the second method proposed by Lu et al.

碩士
國立臺灣大學
國際企業學研究所
104
According to domestic and foreign researches, stop-loss strategies can prevent investors occurring disposition effect, that is holding their losing investments too long and selling the stocks which are winners early. However, the most important thing to investors is return. Therefore, this thesis examine whether stop-loss strategy can lead to higher return or reduce losses for investors. Different from the past research, this thesis uses all the stocks in Taiwan stock market from 2004 to 2014. To avoid the difference of volatility causes different degree of difficulties to reach the stop price, this thesis sets the stop price based on every stock’s daily return standard deviations. In order to prevent other factors making effects on the result, this thesis does not set any specific condition and signal to buy the stocks. In addition, this thesis uses regression to examine whether stock characteristics can affect the validity of stop-loss strategies. In the last we examine whether stop-loss strategy can reduce investment risk. The result shows that stop-loss strategies neither reduce the losses nor increase the return, and investors cannot judge the validity of stop-loss strategies by specific stock characteristic. However, stop-loss strategies are able to reduce investment risk for investors significantly, avoid investors having a huge loss in one investment.

碩士
東吳大學
經濟學系
105
In view of the absolute returns and rewards achieved through stop-loss mechanisms adopted by majority of research institutes, the objective of this study aims to base the research on avoiding the downward share price risks with technical analysis and reducing investment risks from further declines with simple stop-loss mechanism, rather than attempts to create rewards. This study reviews the longest back-testable periods of major country indexes, ranging from 19~67 years, and tests on individual market’s new-low breaking points. The results suggest 60.58%~68.25% chance of a further fall after breaking previous low among the countries studied, and indicate over 60% and above chance to reduce loss with the stop-loss mechanism.

碩士
國立高雄應用科技大學
金融資訊研究所
101
Empirical study aims to establish straightforward investment strategy, and can effectively reduce risk, improve profitability. Research data during 1/1/2002 until 12/31/2011, a total of 2487 pen day backtesting data. First select the front nine of Taiwan stocks traded price as the underlying stock, then choose technical indicators - Moving Average gold cross or death cross as buying and selling timing, with One-stage or two-stage stop profits, stop loss condition as exit timing, there are 42 kinds of strategies empirical results to analyze difference relative to the market buy and hold strategies, and based on "enter strategy" and "exit strategy" investment portfolios to research performance and suitability. The empirical results obtained the following conclusions: First, the use of appropriate stop profits and stop loss parameters as the exit strategy of the performance is better than no set the parameters, which can effectively raise profitability. Second, the use of MA-day parameters should be short (eg MA5_MA20) not too long, can effectively improve profit performance. Third, the establishment of a stop profits and stop loss parameters in approximately 20% to 30%, and obtain better returns performance. Fourth, the entry strategy grouping composed of three portfolios, all returns performance are better than TAIEX, in which the investment portfolio "MA5_MA20" is best performance. Fifth, the exit strategy grouping composed of 14 portfolios, many of which return performance is better than the broader market, Sharpe performance are all better than the broader market. The empirical results, "two-stage stop profits, stop loss" exit strategy group formed portfolios in raising the performance has significant benefits.

碩士
國立臺灣大學
財務金融學研究所
96
This paper investigates different types of investors to see those who would use the “stop-loss” and “stop-gain” mechanisms, and how they affect their behavior. Furthermore, whether investors execute the mechanism or not is also an important factor concerning the effectiveness of those two mechanisms. This study also looks at the “disposition effect”, by which setting stop-loss could prevent investors from holding the losers too long. The main findings are as following: 1. Highly-educated or high-income investors would be more likely to use the stop-loss order or stop-gain order. 2. Contrary to the usual understanding, we find a “reverse disposition effect” exist in our sample, however, if we looked at only the investors who set the stop-loss point higher than stop-gain point, the disposition effect will indeed appear. 3. There are more investors who will postpone executing the stop-loss orders than stop-gain orders. 4. Investors who set higher stop-loss point or lower stop-gain point would be more likely to face disposition effect. 5. Investors who use the stop-loss or stop-gain mechanisms would have higher return, higher sell-turnover and lower holding period. 6. Comparing to investors who didn’t execute the orders duly, those who execute the orders around the point (±3%) they had set would make the highest return. Comparatively, the behavior of disposition effect would result in the lowest return.

碩士
高苑科技大學
經營管理研究所
107
In recent years, we have seen the rise of foreign exchange margin trading. But generally speaking, investors in the trading market often neglect the importance of setting stop loss and do not follow any standard in setting the scope of stop loss rationally so that it is impossible to rationally determine the stop loss point, which ultimately leads to serious losses of funds. Consequently, the main purpose of this study is to find the most appropriate stop loss mechanism. This study used, however, the trading system of MetaTrader4 as platform and employed various technical index trading strategies to conduct statistical analyses of the data collected from 1 September 2017 to 31 August 2018. This study compared, therefore, the three currencies of the euro against the US dollar, the British pound against the US dollar and the US dollar against the Japanese currency to establish the stop loss range in foreignexchange trading. The results of empirical analysis showed that the optimal stop loss range for each time period and cycle was as follows: (1) For EUR/USD: the one-hour period (H1) with 06, 12 as cycles fell in the price range of 400-500 USD, and the four-hour period (H4) fell in the price range of 500-600 USD, the H4 with 12 as cycle fell in the range of 600 to 700 USD; (2) For GBP/USD: the H1 with 06 as cycle fell between 500 and 600 USD, the H1with 12 as cycle fell between 400 and 500 USD, while the H4 with 06, 12 as cycles fell between 600 and 700 USD; (3) For USD/JPY: the H1 with 06, 12 as cycles fell between 300 and 400 USD, and the H4 with the cycles of 06 and 12 fell between 500 and 600 USD. According to the stop loss mechanism shown above, we could, therefore, help investors to find the optimal range to reduce unpredictable risks.

碩士
亞洲大學
資訊工程學系碩士班
101
When compared to unipolar devices Insulated gate bipolar transistors (IGBT) have a significant lower forward (on-state) voltage drop, due to the high level injection of minority carrier from the anode. However this is at the expense of higher energy losses and a larger turn-off time (switching loss). The on-state/switching trade-off performance of the IGBT is strongly dependent on the anode injection efficiency. The punch through- IGBT (PT-IGBT) offers a better on-state/switching trade-off. To obtain at the same time a low on-state voltage drop, fast switching speed and high short-circuit ability, the anode injection of minority carrier should be enhanced during on-state and weakened or ideally eliminated during turn-off. A possible way to achieve this is to use a low doped thin buffer layer (Field stop), in which the minority carriers need to be removed partly or completely before the IGBT starts to be turned-off. The performance of IGBTs has been improved significantly by using a trench-gate structure similar to that used in trench MOSFETs. The trench IGBT is a MOS-bipolar structure in which the channel is formed on the side walls of a vertical groove. This structure allows a reduction of the on-state voltage drop by suppressing the JFET resistance, which results from the constriction of the current flow in the region between adjacent cells in the planar structure. Furthermore, the vertical channel requires less chip area and permits an increase in cell density [5]. In the forward conduction state of the conventional planar IGBTs, current crowding and electric field curvature in the N-base region under the gate electrode, sandwiched between two neighboring P-base wells (JFET region) does not allow sufficient conductivity modulation due to both hole and electron currents. This effect is alleviated in trench IGBTs. The use of a trench gate structure allows enhanced excess carrier injection on the emitter side, between p-base wells, promoting conductivity modulation, leading to a lower Vce(sat). Trench structure also provides a more robust IGBT as it more effectively suppresses the latch-up of the PNPN four layer device. This is due to the fact that the N+ emitter can be made narrower and the P-base can be more heavily doped. Suppression of PNPN latch-up and control of IGBT saturation current are used in IGBTs to provide products with excellent short circuit withstand capability and robust FBSOA. The IGBT has an easy gate drive, and faster switching speed. However, this switching speed is limited by the existence of an inherent current tail, which increases the switching losses. Various approaches have been proposed to optimize switching losses, which is often achieved at the expense of forward voltage drop. This switching loss can be minimized by anode engineering. Although there have been several anode engineering schemes in the past, the main objective of anode engineering to obtain an efficient tradeoff between the forward voltage drop and turn-off. This thesis shows the simulation of an IGBT structure and compares its characteristics with the conventional device. A trench gate field stop PT-IGBT has been designed using Synopsis T-CAD simulation tool. The work described in this thesis studies the behavior of 1200V IGBT with low forward voltage dropof 2.1 V at 140A/cm2, the variations in the device characteristics and its optimization, and improvements in switching characteristics.

碩士
東吳大學
財務工程與精算數學系
104
By the internationalization of the RMB, the exchange rate of USD against RMB fluctuates bilaterally, the RMB no longer unilaterally appreciates as before. The domestic manufacturer who bet on the appreciation of RMB, and purchase the RMB-related derivatives such as target redeemable forward contracts (TRF) therefore suffers. This paper investigates the TRF issued by the E-SUN bank in Taiwan, and employing the Monte Carlo simulation for the valuation and risk analysis of TRF. Also this paper uses regression for analyzing the sensitivity of the value and VaR (value at risk) of TRF to key parameters. This paper finds that the value of TRF with stop-loss mechanism is higher than that of TRF without stop-loss mechanism, moreover, the VaR of TRF with stop-loss mechanism is lower than that of TRF without stop-loss mechanism. Keywords: Barrier option, Monte Carlo simulation, Structured Note, Value at Risk, Target Redemption Forward.

碩士
銘傳大學
資訊管理學系碩士在職專班
99
In the long-term investment tools, mutual fund and stock are the two most popular objectives, and dollar cost averaging investment fund is widely taken as an object of investment .This study discusses if investors make a difference comparation in investing Taiwan Top 50 constituent stock and mutual fund in the way of dollar cost averaging, so as to decide which is a better choice of the return on investment, mutual fund or Taiwan Top 50 constituent stock invested in the way of dollar cost averaging. Furthermore, leading stocks in each industry are selected in Taiwan Top 50 constituent stock as the research objective, except the leading stocks and equity funds in finance and opto-electronics industry. They are divided into short term (three years), medium term (five years) and long term (eight years) for mock trading and analysis of annualized return. The results show that no matter data of short term, medium term or long term, there are no significant differences of annualized return between the leading stocks except those in finance and opto-electronics industry and industry leading stocks; There are also no significant differences of annualized return between industry leading stocks and equity funds; but there are significant differences of annualized return between the leading stocks except those in finance and opto-electronics industry and stock funds. The results show that the annualized return of the leading stocks except those in finance and opto-electronics industry is better than that of industry leading stocks and equity funds. By the results of the study, references for investors in the way of dollar cost averaging during the long-term investment are intended to be provided.

碩士
東吳大學
財務工程與精算數學系
104
As the RMB depreciates recently, many investors suffers from the purchase of target redeemable forward contracts (TRF). This paper investigates the TRF issued by the Bank SinoPac in Taiwan, and employing the Monte Carlo simulation for the pricing and risk analysis of TRF. Also this paper uses regression for analyzing the sensitivity of the value and VaR (value at risk) of TRF to key parameters. This paper also considers the impacts of stop-loss condition to the pricing and risk of TRF. This paper finds that the price of TRF with stop-loss condition is higher than that of TRF without stop-loss condition, on the other hand, the VaR of TRF with stop-loss condition is lower than that of TRF without stop-loss condition. This paper also finds that many parameters influence the price and VaR of TRF significantly. Keywords: Monte Carlo simulation, Structured Note, Value at Risk, Target Redemption Forward.