Dissertations / Theses on the topic 'Test handling'

The unit RTH, Radio Access Network Transmission and Home at Ericsson site in Linkoping uses today different test tool which are divided between different projects. Today they do not have any optimal solution for in an easy way execute the test cases from different projects concurrently, and share the test tools between the these projects.

All the execution of test cases which a test tool is needed needs to be configured and started manually which cost both time and money. Since the test tools are very expensive to use, it is desirable to increase the utilization. The purpose of this thesis is to provide RTH a working prototype which can in an intelligent way schedule and then automatically execute the test cases. The prototype shall consist of a web user interface and a scheduler part. The web user interface is going to be that part which the user works in, and the scheduler handles the prioritization and make sure that the test cases are executed.

To reach the goal with a working prototype, PHP and Java were picked as framework for the prototype. The theory behind these programming languages and more can be read in the theory chapter, and all the different methods which were used. The result of the working process can be read in the Result chapter.

The end prototype fulfills the customer’s requirement. Improvement and new functionalities are given as suggestion at the end of this thesis, where we also discuss the working process.

A sociedade está cada vez mais dependente de sistemas embarcados, sendo que na grande maioria das vezes eles operam de maneira invisível aos seus usuários. Essa dependência torna esses usuários vulneráveis a riscos, devido às falhas que podem ocorrer. Essas falhas podem provocar perdas de vidas ou sérios danos materiais e financeiros. Devido a estes fatos, a qualidade destes produtos torna-se um ponto essencial para se ter um sistema estável, livre de erros e com todas as suas funcionalidades sendo executadas. De encontro a isso, a etapa de teste apresenta-se como indispensável e de relevada importância para a obtenção de um produto com uma boa qualidade. Devido ao alto custo de produção e energia gasto com testes, surge a necessidade de novos estudos, sobre diversificados métodos, para se testar um sistema embarcado. Neste contexto, este trabalho tem como objetivo apresentar os estudos iniciais do teste de um Sistema Operacional Embarcado (SOE), através de um estudo de caso focado na rotina de tratamento de exceção do eCos (Embedded Configurable Operating System), pois ela apresenta uma forte interação entre software e hardware, sendo que esta interação é um dos principais desafios encontrados no teste de um software embarcado. Com isso, este trabalho pretende dar o passo inicial para pesquisas relacionadas aos testes de um Sistema Operacional Embarcado. Após a análise dos experimentos, pôde-se notar que a principal característica do Sistema Operacional Embarcado eCos, a configurabilidade, é um ponto de dificuldade extra para a realização dos testes, pois exige um estudo detalhado do código do SOE, o qual é totalmente genérico, antes do planejamento dos testes, podendo ser gasto muito tempo nessa atividade. Outro ponto é que o teste torna-se totalmente dependente do hardware. Entretanto, os resultados experimentais apresentados para o estudo de caso do presente trabalho foram satisfatórios.
Society is increasingly dependent on embedded systems, which in most cases operate in an invisible manner to its users. This dependence makes the user vulnerable to risks due to failures that may occur. These failures can cause loss of lives or serious property and financial damage. Because of these facts, the quality of these products becomes a key point to have a stable system, free of errors and with all the features running. This testing is of essential importance to obtain a product with good quality. Due to the high cost of production and energy spent on tests, there is a need for further studies on different methods, to test an embedded system. In this context, this work aims at presenting the initial studies as the testing of the Embedded Operating System. The case study was focused on the exception handling routine of the eCos (Embedded Configurable Operating System), because it has a strong interaction between software and hardware, and this interaction is one of the main challenges encountered in testing embedded software. Therefore, this work aims at taking the first steps towards research related to testing an Embedded Operating System. After analyzing the experiments, it was noted that the main feature of the Embedded Operating System, eCos, the configurability, is an extra point of difficulty for the tests. It requires a detailed study of the code eCos, which is completely general, before the planning of tests, and could be spent much time in this activity. Another point is that the test becomes totally dependent on hardware. However, the experimental results presented for the case study of this study showed satisfactory.

Trouble Reports (TRs) encompass the details of a fault-ended test scenario during Radio Base Stations (RBS) verification phase. TR handling is a daily routine for designers in Ericsson Radio Software Department which has been reversed into a bottleneck. TR handling is a log-centric procedure and getting right information from log files is not easily applicable due to various reasons. In this research, we went first to pinpoint those barriers which put difficulty in TR handling procedure. This phase accomplished with comments from TR handlers combined with standard principles for fault handling. A list of hindrances is recognized by the end of this part of project which introduced the main area we should work for performing improvements in TR handling. In second phase of project, we concentrated on cryptic data as a major barrier in log analysis and consequently in TR handling. Cryptic data refers to hardware state during various intervals which pass useful information about hardware in different settings. Troubleshooters can not easily receive the message from this kind of data due to its low level essence. A solution is proposed to handle this difficulty by helping users with the final message which is passed through cryptic data. We get test settings from log file and try to make an expectation on hardware state during test routine, then we make a comparison between what is expected and what is observed in log file. Any inconvenience between observed value and expected one might indicate a deficiency in relevant blocks. Finding out the expected value in each step is done through two different solutions. At first solution, we dig into calculation details in each block and try to perform all steps in parallel with application but in a safe-mode to make sure that nothing might affect the results. For second solution, we do not engage with calculation complexity. Some test cases are designed to target selected blocks in application and watch out the behavior in any possible configuration. We have a database from all possible states which have impact on a block’s result based on that block’s characteristics. Functional test is run on each Radio separately and presented information in a TR should be compared with recorded data for that specific Radio. These solutions are implemented as a web application which receives Trace and Error log as the source for catching test settings. Then correctness of presented data might be confirmed by recalculation or reading from results of functional test.

ISO 26262 is an international standard about functional safety published on 2011, aiming to avoid failure in safety related electrical and electronic systems in passenger cars. A corresponding standard for heavy vehicles is expected to be published in a few years’ time. In order to be well-prepared, the heavy vehicle manufacturer Scania decides to start investigating what impact ISO26262 can bring. At Scania, modularization is one of the most important features of the product, which means several modules can be combined together into a vehicle in a variety of ways, so that highly configurable products can be provided for the customer. Huge number of unique module combinations bring big challenges to systems integration testing department REST/I in Scania because of limited time and resource availability for testing. Nowadays, people in REST/I deal with the variant mainly based on human experience, which is quite difficult to obtain the exact complete variant information and concrete testing coverage. In order to fulfill the requirement related with variant handling in ISO26262, better variant handling methods are proposed in this thesis, which can mainly be divided into two parts: method for variant generating and method for configuration selecting. To simplify the implementation work of this thesis, only the ECU components are focused on (other components such as sensors are ignored), and the risk-based feature is not added to the configuration selecting. Variant generating is to generate variant information from Allocation Element Diagram in Sesamm database systematically. According to the generated variant information, the configuration can be selected automatically by using Greedy best-first-search algorithm based on the proposed testing coverage metrics. Since all these work can be done automatically by computer, REST/I not only can work more efficiently by saving a lot of labor resource, but also can avoid mistakes caused by anthropogenic factors. However, not all the data needed for the automation are existed today, so the suggestions for consummation of the data to be ready for implementing the proposed methods are also mentioned in this thesis.

Many powders and particulate solids are stored and handled in large quantities across various industries. These solids often encounter handling and storage difficulties that are caused by the material cohesion. The cohesive strength of a bulk material is a function of its past consolidation stress. For example, high material cohesive strength as a result from high storage stresses in a silo can cause ratholing problems during discharge. Therefore, it is essential to consider the stress-history dependence when evaluating such handling behaviour. In recent years the Discrete Element Method (DEM) has been used extensively to study the complex behaviour of granular materials. Whilst extensive DEM studies have been performed on cohesionless solids, much less work exists on modelling of cohesive solids. The commonly used DEM models to model adhesion such as the JKR, DMT and linear cohesion models have been shown to have difficulty in predicting the stress-history dependent behaviour for cohesive solids. DEM modelling of cohesive solid at individual particle level is very challenging. To apply the model at single particle level accurately would require one to determine the model parameters at particle level and consider the enormous complexity of interfacial interaction. Additionally it is computationally prohibitive to model each and every individual particle and cohesion arising from several different phenomena. In this study an adhesive elasto-plastic contact model for the mesoscopic discrete element method (DEM) with three dimensional non-spherical particles is proposed with the aim of achieving quantitative predictions of cohesive powder flowability. Simulations have been performed for uniaxial consolidation followed by unconfined compression to failure using this model. Additionally, the scaling laws necessary to produce scale independent predictions for cohesionless and cohesive solids was also investigated. The influence of DEM input parameters and model implementation have been explored to study the effect of particle (meso-scale) properties on the bulk behaviour in uniaxial test simulation. The DEM model calibration was achieved using the Edinburgh Powder Tester (EPT) – an extended uniaxial tester to measure flowability of bulk solids. The EPT produced highly repeatable flowability measurements and was shown to be a good candidate for DEM model calibration. The implemented contact model has been shown to be capable of predicting the experimental flow function (unconfined compressive strength versus the prior consolidation stress) for a limestone powder which has been selected as a reference solid in the Europe wide PARDEM research network. Contact plasticity in the model is shown to affect the flowability significantly and is thus essential for producing satisfactory computations of the behaviour of a cohesive granular material. The model predicted a linear relationship between a normalized unconfined compressive strength and the product of coordination number and solid fraction. Significantly, it has been found that contribution of adhesive force to the limiting friction has a significant effect on bulk unconfined strength. Failure to include the adhesive contribution in the calculation of the frictional resistance may lead to under-prediction of unconfined strength and incorrect failure mode. The results provide new insights and propose a micromechanical based measure for characterising the strength and flowability of cohesive granular materials. Scaling of DEM input parameters in a 3D simulation of the loading regimes in a uniaxial test indicated that whilst both normal and tangential contact stiffness (loading, unloading, and load dependent) scales linearly with radius of the particle, the adhesive forces scales with the square of the radius of the particles. This is a first step towards a mesoscopic representation of a cohesive powder that is phenomenological based to produce the key bulk characteristics of a granular solid and the results indicate that it has potential to gain considerable computational advantage for large scale DEM simulations. The contact model parameters explored include particle contact normal loading stiffness, tangential stiffness, and contact friction coefficient. The DEM model implementation parameters included numerical time step, strain rate, and boundary condition. Many useful observations have been made with significant implications for the relative importance of the DEM input parameters. Finally the calibration procedure was applied to a spray dried detergent powder and the simulation results are compared to whole spectrum of loading regime in a uniaxial experiment. The experimental and simulation results were found to be in reasonable agreement for the flow function and compression behaviour.

Individuals often have difficulty identifying the heuristics that affect their decision-making (NOFSINGER and BAKER, 2002). However, Menkhoff, Brozynki and Schmidt (2006) show that individuals knowledgeable about the influence of heuristics in decision making may have fewer episodes of bounded rationality compared to individuals unrelated to this knowledge. According to Mussweiler et al. (2000), the Anchorage is one of the most striking influences on judgment and decision making. In this context, the main objective of this research was to verify the existence of heuristics in the anchoring of the numerical estimates of experts and non-specialists in real estate, when subjected to experimental tasks related to the field of domain expert and outside their field of knowledge, in order Thorsteinson et al. (2008), who argue that the effects of Anchorage makers are reduced when subjects have more knowledge about the problems in question. Thus, we performed a experiment with 324 subjects makers were divided into groups for Calibration and Experimental Groups, through the application of two decision-making tasks (T1 and T2), using the model Jacowitz and Kahneman (1995). It is noteworthy that many studies have been conducted in Anchorage to verify the presence of the anchoring heuristic in the process of judgment and decision making. However, this study, besides the proposal to study the effects of Anchorage during the decision making process related to quantitative, as the main contribution presented additional verification of the manifestation of Anchorage as a control variable (knowledge) is removed from the decisionmaking task. This procedure is called by Cozby (2006) a test handling, consisting in an attempt to directly measure whether the manipulation of a variable has the intended effect on the respondents. The results obtained through statistical t test for independent samples showed that there is evidence of the anchoring heuristic in decision making of individual members of the experimental group of non-experts (anchor and anchor low-high) in both T1 and T2. With respect to experts, it was revealed that T1 is no evidence of the anchoring heuristic, which corroborates the results found by Jacowitz and Kahneman (1995) and Luppe (2006). However, in relation to the second decision-making task (T2), it was observed that the mean differences were not statistically significant, implying that, at T2, are not found evidence of Anchorage in the decision process of experts, contradicting the results by Norfhcraft and Neale (1987) and Dorow (2009), who found evidence of Anchorage in the decision process when they were real estate brokers who underwent estimates for the real estate market.
Os indivíduos normalmente têm dificuldade em identificar as heurísticas que afetam suas tomadas de decisões (BAKER e NOFSINGER, 2002). No entanto, Menkhoff, Schmidt e Brozynki (2006) revelam que indivíduos conhecedores da influência das heurísticas no processo decisório podem apresentar menos episódios de racionalidade limitada se comparados aos indivíduos alheios a esse conhecimento. Segundo Mussweiler et al (2000), a Ancoragem consiste em uma das influências mais notáveis em julgamento e tomada de decisão. Neste contexto, o objetivo principal desta pesquisa consistiu em verificar a existência da heurística da Ancoragem nas estimativas numéricas de especialistas e não-especialistas em mercado imobiliário, quando submetidos a tarefas experimentais relacionadas ao campo de domínio do especialista e fora do seu campo de conhecimento, tendo em vista Thorsteinson et al. (2008), os quais argumentam que os efeitos da Ancoragem são reduzidos quando sujeitos decisores têm mais conhecimento acerca dos problemas em questão. Assim, foi realizado um quase-experimento com 324 sujeitos decisores, divididos em Grupos de Calibragem e Grupos Experimentais, por meio da aplicação de duas tarefas decisórias (T1 e T2), utilizando-se o Modelo de Jacowitz e Kahneman (1995). Ressalta-se que muitos estudos sobre Ancoragem têm sido realizados no sentido de verificar a presença da heurística da Ancoragem no processo de julgamento e tomada de decisão. No entanto, este estudo, além da proposta de estudar os efeitos da Ancoragem durante o processo decisório relacionado a decisões quantitativas, apresentou como principal contribuição adicional a verificação da manifestação da Ancoragem quando uma variável de controle (conhecimento) é retirada da tarefa decisória. Tal procedimento é denominado por Cozby (2006) como um Teste de Manipulação, consistindo em uma tentativa para medir diretamente se a manipulação de uma variável tem o efeito pretendido sobre os respondentes. Os resultados obtidos, por meio do teste estatístico t para amostras independentes, demonstraram que há evidências da heurística da Ancoragem no processo decisório dos indivíduos integrantes do Grupo Experimental de não-especialistas (âncora baixa e âncora alta), tanto em T1 quanto em T2. No que tange aos especialistas, foi possível perceber que em T1 há evidências da heurística da Ancoragem, corroborando com os resultados encontrados por Jacowitz e Kahneman (1995) e Luppe (2006). No entanto, em relação à segunda tarefa decisória (T2), observou-se que as diferenças de médias não foram estatisticamente significantes, inferindo-se que, em T2, não são encontradas evidências da Ancoragem no processo decisório dos especialistas, contrariando os resultados encontrados por Norfhcraft e Neale (1987) e Dorow (2009), os quais encontraram evidências da Ancoragem no processo decisório de corretores imobiliários quando foram submetidos à realização de estimativas referentes ao mercado de imóveis.

The relationships between and among questions from a survey in three different areas, namely food and food related health, food safety and hygiene, and information gathering concerning food, with the main focus on the first two areas, are investigated. Data from older people have been analyzed. The questions are both about knowledge and behavior in relation to the areas mentioned above. Earlier studies have shown that people in general do not know about or act according to dietary advises about health and hygiene. The results from this study show that there are fairly strong correlations for all groups but men when it comes to knowledge and behavior about fruit and vegetable consumption. The results also show a significant association between fruit and vegetable consumption and gender and that women consume more than men. A factor model estimating the relationships between health, hygiene and information shows that the fitted factor model can be considered having an acceptable fit, with at least two variables, knowledge about the recommended amount of fruit and vegetables to eat per day and trust for grocery stores concerning food and health, being classified as good indicators.

Early life events, in both humans and animals, may influence on brain development, leading to long-term consequences on adult behavior. Thus, this study was carried out in order to assess the influence of neonatal tactile stimulation (TS), when associated to sub-therapeutic administration of sertraline (SERT) in rats, on behavioral parameters related to depression-like. We also evaluated the influence of the TS-SERT association on the hypothalamic-pituitary-adrenal axis, as well as on the development of anxiety-like symptoms, assessed in different behavioral. Male rat pups were daily submitted to 10 min TS, from postnatal day (PND) 8 to PND 14. On DPN50, rats were first exposed to forced swimming test (FST). Then, after 24h, it was conducted the SERT administration and, 30 minutes later, animals were submitted to new evaluation on FST. On DPN51, rats were evaluated on the elevated plus maze (EPM). The animals were euthanized 24h after the behavioral tasks, followed by biochemical assessments. TS per se decreased depression behaviors, since there was an increase on swimming time and a decrease on immobility time in FST. On the EPM task, TS diminished behaviors related to anxiety, as shown in the increase of time on open arms and decrease on anxiety rates. In addition, TS also decreased corticosterone and cortisol plasma levels. All these results were enhanced when TS was associated to a sub-therapeutic dose of SERT. From these results, we suggest that neonatal handling TS may accomplish beneficial influence on parameters of depression, promoting the ability to cope with stressful situations in adulthood.
Eventos ocorridos em estágios iniciais do desenvolvimento, tanto em humanos quanto em animais, influenciam o desenvolvimento do cérebro e podem ter consequências em longo prazo na idade adulta. Assim, o presente estudo foi realizado para avaliar a influência da estimulação tátil neonatal (ET), associada à administração de uma dose subterapêutica de sertralina (SERT), sobre parâmetros comportamentais relacionados a depressão em ratos. A influência da associação ET-SERT sobre o eixo hipotálamo-pituitária-adrenal, assim como sobre o desenvolvimento de comportamentos de ansiedade foram também avaliados. Filhotes machos de ratos Wistar foram diariamente submetidos à ET durante 10 min, desde o dia pós-natal (DPN) 8 até o DPN14. No DPN50, os animais foram previamente expostos ao nado forçado. Após 24h, receberam a administração de SERT e, decorridos 30 minutos, foram novamente avaliados mediante o estresse do nado forçado. No DPN 51, os animais também foram avaliados no labirinto em cruz elevado (LCE). Os animais foram eutanasiados 24h após os testes comportamentais, seguindo-se as avaliações bioquímicas. A ET per se reduziu comportamentos de depressão, tendo em vista a observação do maior tempo de nado e menor tempo de imobilidade no teste do nado forçado. No teste do LCE, a ET diminuiu comportamentos relacionados à ansiedade, como observado pelo maior tempo nos braços abertos e pelo menor índice de ansiedade. Além disso, animais expostos à ET apresentaram menores níveis plasmáticos de corticosterona e cortisol. Todos estes resultados foram potencializados quando a ET foi associada a uma dose subterapêutica de SERT. Considerando estes dados, é possível propor que a ET neonatal é capaz de exercer influências benéficas sobre parâmetros de depressão, favorecendo a capacidade para lidar com situações estressantes na vida adulta.

Excessive and compulsive ethanol drinking is one of the most serious public health issues. Therefore, it is vital to increase the knowledge about risks and protection for alcohol use disorders (AUD) to optimize prevention and treatment strategies. Ethanol consumption commonly initiates during adolescence when extensive neuronal maturation and development also occurs. Early exposure to ethanol is a risk factor for AUD, but the effects of adolescent drinking and the basis for the individual susceptibility to AUD are not fully understood. The interactions between genotype and environmental factors determine the individual risk for AUD and this thesis aimed to examine the environmental impact. The specific aims were to investigate 1) how early-life conditions affect adolescent voluntary ethanol drinking, behavioural profiles, endogenous opioids and response to treatment with an opioid antagonist (naltrexone), and 2) whether alterations detected in the offspring may be mediated by variations in maternal behaviour. A rodent maternal separation (MS) model was used to mimic a protective and risk-inducing early-life environment, respectively, with the use of 15 min (MS15) or 360 min (MS360) of daily MS. The main findings were 1) the MS360, but not the MS15 rats, responded to naltrexone following adolescent ethanol drinking; all adolescent rats had a high voluntary ethanol intake independent of early environmental conditions whereas in the adult groups the MS360, but not the MS15 rats, increased their ethanol intake and preference over time; adolescent ethanol exposure resulted in higher dynorphin levels in hippocampus and higher Met-enkephalin-Arg6Phe7 in the amygdala, independently of rearing conditions, 2) behavioural profiling using the multivariate concentric square field™ test showed: the young MS360 rats had increased risk assessment and risk taking behaviour compared to the young MS15 rats; the young MS15 rats increased, whereas the young MS360 rats decreased, their risk assessment and risk taking behaviour over time; differences in pup-retrieval strategies where the MS360 dams retrieved some pups into a safe area but as compared to MS15 rats they left more pups in a risk area; increased risk assessment behaviour in the MS360 dams immediately after weaning. Taken together, early-life environmental conditions alter adult but not adolescent drinking, the response to naltrexone, and behaviour in dams and offspring. Adolescent rats consumed more ethanol independent of rearing conditions and displayed increased opioid levels in brain areas related to cognition and addiction.

This thesis deals with the rebound hammer method as a tool for approximation of the time limit for handling the concrete. The theoretical part is focused on three fields of knowledge - rebound hammer test, production of precast concrete components and statistical data analysis. The following practical part deals with the design of two single-parameter linear functions for two types of rebound hammer testers - SilverSchmidt L and SchmidtOriginal N. Statistical data processing is completed by the analysis of influential points by using the Cook’s distance. The resulting statistical models are compared with commonly used relationships.

In this essay I have examined text, speech and action in Chekhov's Three sisters, as well as how the characters express their wills in action and in speech. I have used different methods such as action analysis, the five W questions and BSI (background, situation and intention). These methods are similar to those actors use to uncover situatedness of characters and to find motivated and intentional patterns behind their actions. I have used these techniques to identify the plot, explore what happened before the plot begins and to read the text in terms of action. The aim was also to find the counter points where wills expressed in action are contradicted by wills expressed in speech. Using these methods I found that the characters in Three sisters manage their lives dreaming about the future or recalling the past, except the character Natasha. Her goal is obvious, she runs her will with clear intentions and she uses functional means to achieve her goal. The main conclusion is that the characters’ wills expressed in speech, are contradicted by will expressed in action. This creates a discrepancy. The characters are controlled by their unspoken desires which drive them further away from their expressed goals. It seems to be paradoxical, but it makes perfect sense since it drives the plot forward, deepens the image of the characters and the relationships between them. It also becomes clear when actions are interpreted from wills and motivations, stated or unstated, and I argue that action is not only the doing but also the intention to do; this makes sense when events in a dramatic text are analyzed as relational and conditional.

This thesis aimed to develop a paradigm for the study of anomaly handling and to investigate the factors that influence success in detecting and classifying anomalies. A simulated anomaly-handling environment was created to mimic an intelligence analysis task in a security setting. A series of experiments was designed to test hypotheses concerning sources of difficulty in detecting potential anomalies and making decisions about appropriate classifications of potentially anomalous events. Results across all experiments showed that complex problems, representing anomalies, were more difficult to solve than simple problems, and that this poor performance was consistent with the use of suboptimal strategies based on recognition of perceptual characteristics rather than inferences drawn from available data. Performance on complex problems was reduced still further when participants were exposed to trials that established a mental set. However, performance was improved when participants were given feedback on the correctness of their responses to each trial, which eliminated the negative effects of exposure to mental set. Another factor that impacted on successful decision-making was the cost of making errors. When participants were faced with a penalty for making incorrect decisions, solution rates improved compared with when performance was not related to reward. This has consequences for anomaly handling industries where the consequence of failure of often high. Unexpectedly, a number of the results indicated that there are situations where mental set may confer a benefit to decision making in a task of anomaly categorisation. Given the dominance of recognition-based strategies, it appears that mental set can refine the detection of perceptually relevant patterns, which can signal sudden changes in pattern that can lead to a switch from recognition-based to inferential task solution strategies. Overall, the merits for the use of simulated environments in critical decision making areas are discussed, and the contributory factors towards successful anomaly handling are analysed.

Software companies spend a lot of resources on debugging, a process where previous solutions can help in solving current problems. The bug tickets, containing this information, are often time-consuming to read. To minimize the time spent on debugging and to make sure that the knowledge from prior solutions is kept in the company, an evaluation was made to see if summaries could make this process more efficient. Abstractive and extractive summarization models were tested for this task and fine-tuning of the bert-extractive-summarizer was performed. The model-generated summaries were compared in terms of perceived quality, speed, similarity to each other, and summarization length. The average description summary contained part of the description needed and the found solution was either well documented or did not answer the problem at all. The fine-tuned extractive model and the abstractive model BART provided good conditions for generating summaries containing all the information needed.
Vid mjukvaruutveckling går mycket resurser åt till felsökning, en process där tidigare lösningar kan hjälpa till att lösa aktuella problem. Det är ofta tidskrävande att läsa felrapporterna som innehåller denna information. För att minimera tiden som läggs på felsökning och säkerställa att kunskap från tidigare lösningar bevaras inom företaget, utvärderades om sammanfattningar skulle kunna effektivisera detta. Abstrakta och extraherande sammanfattningsmodeller testades för uppgiften och en finjustering av bert-extractive- summarizer gjordes. De genererade sammanfattningarna jämfördes i avseende på upplevd kvalitet, genereringshastighet, likhet mellan varandra och sammanfattningslängd. Den genomsnittliga sammanfattningen innehöll delar av den viktigaste informationen och den föreslagna lösningen var antingen väldokumenterad eller besvarade inte problembeskrivningen alls. Den finjusterade BERT och den abstrakta modellen BART visade goda förutsättningar för att generera sammanfattningar innehållande all den viktigaste informationen.

Dans le contexte actuel, il y a un besoin évident de techniques de fouille de textes pour analyser l'énorme quantité de documents textuelles non structurées disponibles sur Internet. Ces données textuelles sont souvent représentées par des matrices creuses (sparses) de grande dimension où les lignes et les colonnes représentent respectivement des documents et des termes. Ainsi, il serait intéressant de regrouper de façon simultanée ces termes et documents en classes homogènes, rendant ainsi cette quantité importante de données plus faciles à manipuler et à interpréter. Les techniques de classification croisée servent justement cet objectif. Bien que plusieurs techniques existantes de co-clustering ont révélé avec succès des blocs homogènes dans plusieurs domaines, ces techniques sont toujours contraintes par la grande dimensionalité et la sparsité caractérisant les matrices documents-termes. En raison de cette sparsité, plusieurs co-clusters sont principalement composés de zéros. Bien que ces derniers soient homogènes, ils ne sont pas pertinents et doivent donc être filtrés en aval pour ne garder que les plus importants. L'objectif de cette thèse est de proposer de nouveaux algorithmes de co-clustering conçus pour tenir compte des problèmes liés à la sparsité mentionnés ci-dessus. Ces algorithmes cherchent une structure diagonale par blocs et permettent directement d'identifier les co-clusters les plus pertinents, ce qui les rend particulièrement efficaces pour le co-clustering de données textuelles. Dans ce contexte, nos contributions peuvent être résumées comme suit: Tout d'abord, nous introduisons et démontrons l'efficacité d'un nouvel algorithme de co-clustering basé sur la maximisation directe de la modularité de graphes. Alors que les algorithmes de co-clustering existants qui se basent sur des critères de graphes utilisent des approximations spectrales, l'algorithme proposé utilise une procédure d'optimisation itérative pour révéler les co-clusters les plus pertinents dans une matrice documents-termes. Par ailleurs, l'optimisation proposée présente l'avantage d'éviter le calcul de vecteurs propres, qui est une tâche rédhibitoire lorsque l'on considère des données de grande dimension. Ceci est une amélioration par rapport aux approches spectrales, où le calcul des vecteurs propres est nécessaire pour effectuer le co-clustering. Dans un second temps, nous utilisons une approche probabiliste pour découvrir des structures en blocs homogènes diagonaux dans des matrices documents-termes. Nous nous appuyons sur des approches de type modèles de mélanges, qui offrent de solides bases théoriques et une grande flexibilité qui permet de découvrir diverses structures de co-clusters. Plus précisément, nous proposons un modèle de blocs latents parcimonieux avec des distributions de Poisson sous contraintes. De façon intéressante, ce modèle comprend la sparsité dans sa formulation, ce qui le rend particulièrement adapté aux données textuelles. En plaçant l'estimation des paramètres de ce modèle dans le cadre du maximum de vraisemblance et du maximum de vraisemblance classifiante, quatre algorithmes de co-clustering ont été proposées, incluant une variante dure, floue, stochastique et une quatrième variante qui tire profit des avantages des variantes floue et stochastique simultanément. Pour finir, nous proposons un nouveau cadre de fouille de textes biomédicaux qui comprend certains algorithmes de co-clustering mentionnés ci-dessus. Ce travail montre la contribution du co-clustering dans une problématique réelle de fouille de textes biomédicaux. Le cadre proposé permet de générer de nouveaux indices sur les résultats retournés par les études d'association pan-génomique (GWAS) en exploitant les abstracts de la base de données PUBMED. (...)
In the current context, there is a clear need for Text Mining techniques to analyse the huge quantity of unstructured text documents available on the Internet. These textual data are often represented by sparse high dimensional matrices where rows and columns represent documents and terms respectively. Thus, it would be worthwhile to simultaneously group these terms and documents into meaningful clusters, making this substantial amount of data easier to handle and interpret. Co-clustering techniques just serve this purpose. Although many existing co-clustering approaches have been successful in revealing homogeneous blocks in several domains, these techniques are still challenged by the high dimensionality and sparsity characteristics exhibited by document-term matrices. Due to this sparsity, several co-clusters are primarily composed of zeros. While homogeneous, these co-clusters are irrelevant and must be filtered out in a post-processing step to keep only the most significant ones. The objective of this thesis is to propose new co-clustering algorithms tailored to take into account these sparsity-related issues. The proposed algorithms seek a block diagonal structure and allow to straightaway identify the most useful co-clusters, which makes them specially effective for the text co-clustering task. Our contributions can be summarized as follows: First, we introduce and demonstrate the effectiveness of a novel co-clustering algorithm based on a direct maximization of graph modularity. While existing graph-based co-clustering algorithms rely on spectral relaxation, the proposed algorithm uses an iterative alternating optimization procedure to reveal the most meaningful co-clusters in a document-term matrix. Moreover, the proposed optimization has the advantage of avoiding the computation of eigenvectors, a task which is prohibitive when considering high dimensional data. This is an improvement over spectral approaches, where the eigenvectors computation is necessary to perform the co-clustering. Second, we use an even more powerful approach to discover block diagonal structures in document-term matrices. We rely on mixture models, which offer strong theoretical foundations and considerable flexibility that makes it possible to uncover various specific cluster structure. More precisely, we propose a rigorous probabilistic model based on the Poisson distribution and the well known Latent Block Model. Interestingly, this model includes the sparsity in its formulation, which makes it particularly effective for text data. Setting the estimate of this model’s parameters under the Maximum Likelihood (ML) and the Classification Maximum Likelihood (CML) approaches, four co-clustering algorithms have been proposed, including a hard, a soft, a stochastic and a fourth algorithm which leverages the benefits of both the soft and stochastic variants, simultaneously. As a last contribution of this thesis, we propose a new biomedical text mining framework that includes some of the above mentioned co-clustering algorithms. This work shows the contribution of co-clustering in a real biomedical text mining problematic. The proposed framework is able to propose new clues about the results of genome wide association studies (GWAS) by mining PUBMED abstracts. This framework has been tested on asthma disease and allowed to assess the strength of associations between asthma genes reported in previous GWAS as well as discover new candidate genes likely associated to asthma. In a nutshell, while several text co-clustering algorithms already exist, their performance can be substantially increased if more appropriate models and algorithms are available. According to the extensive experiments done on several challenging real-world text data sets, we believe that this thesis has served well this objective

Ce travail de thèse concerne l'étude d'extracteurs d'air hybrides éoliens. Il se décompose en des investigations amont et appliquées. Dans la partie amont, des simulations fines ont été effectuées en conduite carrée avec et sans rotation, pour des nombres de Reynolds turbulents de l'ordre de 600, afin d'analyser l'impact de la rotation sur la turbulence. Elles ont montré que la rotation rompt la symétrie de l'écoulement. La partie appliquée est dédiée à la conception d'une nouvelle géométrie d'extracteur d'air. Cette géométrie a été proposée à partir de l'analyse de simulations RANS. Ses performances ont été confirmées par des mesures expérimentales sur banc d'essais. Les tests en soufflerie d'un système de captage d'énergie éolienne, conçu pour l'extracteur, ont mis en évidence son adéquation au régime de fonctionnement de l'extracteur. Les essais expérimentaux de l'extracteur complet, montrent que le système de captage apporte une part significative de l'énergie. Des essais en soufflerie ont permis d'observer le comportement global de l'extracteur
This PhD work concerns the study of hybrid air extractors. It is composed of upstream and applied investigations. In the upstream part, fine simulations are realized in square duct flow with and without rotation to analyse the impact of rotation on turbulence. It is found that rotation removes symmetry property of the flow with turbulent Reynolds number of 600. The applied part is dedicated to the conception of a new air extractor geometry. This geometry is proposed from the analyse of RANS simulations. Its performances are confirmed by experimental measurements on test rig. Wind tunnel tests of a wind power capturing system, designed for the extractor, show a good adequation to the operating regime of the extractor. Experimental investigations on the complete air extractor, show the wind power capturing system brings a significant part of the energy. Wind tunnel tests allow to observe the complete air extractor behaviour

This work studied the information flow and information handling among different parties. The study focused on how educational materials on national tests from The Swedish National Agency for Education is managed by school principals and teachers and how the information is passed on to students. The data material was based on observations and interviews at two different high schools. Six teachers and their classes and principals participated in the study. The Agency was also interviewed. The results showed that opinions on the role of the national test varies between different teachers and the Agency. The study also showed that the dissemination of information in the schools is inadequate, information documents do not reach their target audience, and those who have access to the information do not read enough. This is partly due to lack of follow-up from the responsibles at the schools, but also teachers who do not seek out information that is available themselves. These shortcomings result in teachers carrying out national tests with their classes without being properly briefed, which could have consequences for both the students and the entire test system.

The National Geographic Channel’s miniseries “Snake Salvation” resurrected a vested interest with the heavily documented practices of signs-following believers in central Appalachia. The current body of scholarship surrounding these congregations focuses mostly on oral history narratives and explanations of religious fundamentalism; a critical analysis of the discourse shared by these congregation members is noticeably absent. This thesis explores selected interviews with George Hensley, Andrew Hamblin, Jamie Coots, and Alfred Ball through the interdisciplinary application of discourse analysis paired with social disclosure theory to unveil the underlying struggles with power and personal beliefs expressed by each pastor. The research performed throughout this study spans interviews collected and published from the 1940s to 2014. Through a discourse analysis performed on these interviews coupled with support from sociological and communicative theoretical frameworks, this study looks to highlight struggles with power and authenticity present for signs-following pastors.

The company Ericsson is taking steps towards embracing automating techniques and applying them to their product development cycle. Ericsson wants to apply machine learning techniques to automate the evaluation of a text categorization problem of error reports, or trouble reports (TRs). An excess of 100,000 TRs are handled annually. This thesis presents two possible solutions for solving the routing problems where one technique uses traditional classifiers (Multinomial Naive Bayes and Support Vector Machines) for deciding the route through the company hierarchy where a specific TR belongs. The other solution utilizes a Convolutional Neural Network for translating the TRs into low-dimensional word vectors, or word embeddings, in order to be able to classify what group within the company should be responsible for the handling of the TR. The traditional classifiers achieve up to 83% accuracy and the Convolutional Neural Network achieve up to 71% accuracy in the task of predicting the correct class for a specific TR.

This thesis deals with the analysis of the car driving characteristics depending on the type of drive wheels. The first chapter defines the basic theoretical cars concept as well as procedures for determining the individual variables needed to identify the driving dynamics of vehicles. The practical part describes the plan and the progress of realized measurements on a selected pattern of vehicles, and the measured values are interpreted. The final part deals with the evaluation of the performed measurements and the data identified within these measurements.

O mecanismo de tratamento de exceções disponível em linguagens orientadas a objetos está sendo cada vez mais utilizado e representa cerca de 8% das linhas de código em programas escritos com linguagens de terceira geração, de acordo com algumas pesquisas. No entanto, este mecanismo está entre os menos compreendido pelos desenvolvedores e menos testado. A dificuldade de testar exceções aumenta na fase de teste de integração, porque exceções lançadas em um determinado nível e não tratadas podem subir na hierarquia de chamada. O uso de aspectos também pode introduzir novos tipos de defeitos, como por exemplo um adendo pode inserir lançamentos de exceções no programa sem que exista algum tratamento previsto para elas. Este trabalho apresenta uma proposta para testes de integração estrutural de fluxos de exceções de programas Java e AspectJ. Isto é feito propondo-se alterações na representação do fluxo de exceção no grafo de fluxo de controle integrado e propondo-se novos critérios baseados no fluxo de exceção. Além disso, essas propostas foram implementadas e geraram uma extensão da ferramenta chamada Ja-BUTi/AJ. Para avaliar a proposta deste trabalho, um estudo de caso e um experimento formal foram conduzidos. O estudo de caso teve como objetivo validar a extensão da JaBUTi/AJ desenvolvida e fazer uma avaliação inicial do custo de aplicação dos critérios propostos. O experimento formal realizou uma comparação entre as ferramentas JaBUTi/AJ e VITTAE no contexto do teste do fluxo de exceções. Ambos os estudos são apresentados e discutidos nesta dissertação
Exception handling mechanisms available in object-oriented languages have become increasingly used and account for about 8% of the lines of code in modern programs according to some empirical researches. Yet, this mechanism is among the least understood by programmers and the less tested. The difficulty of testing exceptions increases in the phase of integration testing (e.g. inter-class testing) because exceptions raised on a certain level and not handled can flow through the call hierarchy. Using aspects also cause other types of errors, e.g. an advice can insert exceptions not foreseen by the base program. This paper presents a proposal for structural integration testing of exception flows of Java and AspectJ programs. This is done by proposing changes to the classic control ow graph and by proposing new criteria based on the exception ow. Moreover, these proposals have been implemented in a tool named JaBUTi/AJ. To evaluate the proposal of this work, a case study and a formal experiment have been conducted. The case study had the objective of validating the JaBUTi/AJ extension developed and making a preliminary assessment of the cost to use the tool. The formal experiment made a comparison between the tools JaBUTi/AJ and VITTAE in the context of exception flow test. Both studies are presented and discussed in this dissertation

Esta dissertação tem como objetivo central investigar fatores que moderam a relação entre satisfação pós-recuperação e intenção de recompra no varejo online. Mais especificamente, esta pesquisa visa analisar se os construtos qualidade do relacionamento, resistência à mudança e motivo inferido moderam a relação entre satisfação com o gerenciamento de reclamações e intenção de recompra no varejo online. Para verificar tais moderações, inicialmente foi realizada a etapa exploratória de pesquisa, na qual, por meio de entrevistas em profundidade, evidenciou-se que as variáveis propostas poderiam moderar a relação entre a variável independente e dependente desta dissertação, bem como foi adicionado o construto valor econômico como possível moderador da relação entre satisfação pós-reclamação e intenção de recompra. Em seguida, para testar as hipóteses propostas foi aplicada uma survey junto a consumidores norte-americanos (N=500), operacionalizada com auxílio do Qualtrics, via Mechanical Turk. Como resultado dos testes, tem-se que a qualidade do relacionamento, o motivo inferido e a percepção de valor econômico moderam a relação entre satisfação pós-recuperação e intenção de recompra no varejo online, ou seja, podem explicar, em partes, a não linearidade da relação entre a variável independente e dependente desta pesquisa. Destaca-se ainda que as dimensões da qualidade do relacionamento (a saber: satisfação, confiança e comprometimento) sugeridas neste estudo também moderaram, individualmente, a relação entre a variável independente e dependente desta pesquisa. Sendo assim, esta dissertação contribuiu para a literatura de Marketing e comércio eletrônico ao evidenciar que tanto a qualidade do relacionamento, bem como as dimensões de satisfação, confiança e comprometimento moderam a relação entre satisfação com o gerenciamento de reclamações e intenção de recompra. Além disso, evidencia-se que a elevada percepção de valor econômico, e os motivos inferidos pelos clientes (se positivos), diante da falha inicial ocorrida, também podem sinalizar intenções de recompra dos consumidores junto a sites varejistas.
This dissertation aims to investigate factors that moderate the relationship between satisfaction with complaint handling and repurchase intention in online retail. More specifically, this research aims to examine whether the quality of the relationship constructs, resistance to change and inferred motive moderate the relationship between satisfaction with complaint handling and repurchase intention in online retail. To test such moderations, was taken an exploratory research stage through in-depth interviews in which it became clear that the proposed constructs could moderate the relationship between the independent and dependent variables of this dissertation, as well as the variable economic value was added as a possible moderator of the relationships between satisfaction with complaint handling and repurchase intention. Then, to test the hypotheses was applied a survey with online consumers (N=500) from United States, through Qualtrics and Mechanical Turk. The results indicate that the relationship quality, the inferred motive and the perceived economic value moderated the relationship between satisfaction with complaint handling and repurchase intention in online retail. It is emphasized that the dimensions of relationship quality (satisfaction, trust and commitment) suggested in this study also moderated, in an individual way, the relationship between the independent and dependent variables of this study. Thus, this dissertation has contributed to the literature of marketing and e-commerce by showing that both the relationship quality as well as the dimensions (satisfaction, trust and commitment) moderate the relationship between satisfaction with complaint handling and repurchase intention. Furthermore, it is evident that the high perceived economic value, and the motive inferred by customers (if positive), facing the initial failure occurred, can also indicate repurchase intentions of consumers with online retailers.

Numerosos estudos tem demonstrado a relação entre medidas comportamentais e fisiológicas de estresse em suínos no abate e subseqüente efeitos na qualidade de carne. Uma maneira de estudar esta relação é através de testes comportamentais. O teste Y maze (labirinto em Y) é usado especialmente para determinar relação de aversão frente a diferentes procedimentos de manejo. O objetivo do trabalho foi verificar a eficácia do teste do labirinto em Y na detecção de diferenças comportamentais que permitam categorizar suínos híbridos em grupos de alta e baixa reatividade, e a relação desta com parâmetros fisiológicos e musculares. No ensaio 1 foram verificados dois tipos de labirinto, que diferenciavam-se quanto ao estímulo sonoro, aversivo ou não aversivo, na extremidade final de um dos braços laterais do Y. Quarenta suínos híbrido, F1-Landrace (LD) X Large White (LW), 40 fêmeas, com média de 160 dias de idade e peso médio de 94,1 kg de peso vivo, foram submetidos individualmente ao primeiro teste do labirinto em Y por 3 minutos, em uma área que constava de um corredor principal com 110 cm x 80 cm x 220 cm ( altura x largura x comprimento), e os braços laterais com 110 cm x 80 cm x 180 cm, sem contato visual com o exterior. O experimento identificou as seguintes variáveis como significativas para categorização em dois grupos de reatividade divergente: número de vezes que o animal entra em cada braço (NEB); tempo de permanência no braço principal (TPBP); tempo inicial no braço principal (TIBP); tempo total de permanência em cada braço (TTPB). Não houve efeito de tipo de estímulo sonoro para categorização de reatividade. No segundo experimento, utilizando apenas o estímulo sonoro aversivo, 132 suínos híbridos (LD X LW), sendo 66 machos e 66 fêmeas, foram submetidos ao teste do labirinto Y para obtenção de um subgrupo apresentando reatividade divergente, levando em conta as variáveis de classificação de reatividade definidos no ensaio 1. Quarenta e oito animais (24 fêmeas e 24 machos) foram identificados como animais de alta (AR) ou baixa reatividade (BR) e submetidos a dois manejos no período pré-abate, um tranqüilo (T) e outro estressante (E). Não houve diferença de reatividade entre gêneros (machos e fêmeas). Animais AR e BR apresentaram valores similares (P>0,05) para atividade de creatina quinase plasmática (1193,4 U/L± 218 e 1539,9 U/L±218), lactato plasmático ( 3,91 mmol/L± 0,20 e 4,00 mmol/L± 0,20) e glicogênio muscular 3 e 24 horas após abate (0,12 mg/g ± 0,02, 0,04 mg/g ± 0,01 e 0,08 mg/g±0,02, 0,04±0,01) respectivamente. A reatividade não alterou (P>0,05) a temperatura superficial do animal (AR: 32,9 oC ± 0,46, BR: 32,9 oC ± 0,45), enquanto o manejo E resultou em temperatura superficial de 33,9 oC± 0,45, que superou(P<0,05) o valor de 31,9 o C ± 0,46 observado nos animais do manejo T. Embora os testes tenham apontado para a possibilidade de categorização de reatividade pelo labirinto Y, as medidas fisiológicas examinadas não foram influenciadas pela reatividade divergente, mesmo em situações pré-abate diferenciadas. O manejo pré-abate também não influenciou os resultados das variáveis fisiológicas, exceto para temperatura superficial, que aponta para diferenças na atividade dos animais.
Numerous studies have demonstrated the relationship between behavioral and physiological measures of stress in pigs at slaughter and subsequent effects on meat quality. One way to study this relationship is through behavioral tests. The test Y maze is used especially to determine the relationship of aversion front the different procedures for handling. The objective of this study was to test the effectiveness of Y-maze to detect behavioral differences that allow categorizing crossbred pigs in groups of high and low reactivity, and the relationship between physiological parameters and muscle. In trial I, two types of maze were verified, which were distinguished as the sound stimulus, aversive or not aversive, at the end of one of the side arms Y. Forty pigs crossbreed, F1-Landrace (LD) X Large White (LW), 40 females, with an average of 160 days of age and averaging 94.1 kg body weight were submitted individually to the first test of the Y maze for 3 minutes in an area that consisted of a mean arm with 110 cm x 80 cm x 220 cm (height x width x length), and side arms with 110 cm x 80 cm x 180 cm, without visual contact with the outside. The experiment considered the variables as significant to categorization in two groups of divergent reactivity: number of times the animal enters into each arm (NEA); time of main arm (TMA); time in initial main arm (TIMA) and total time of stay in each arm (TTSA). Had no effect of type sound stimulus to reactivity categorization. The second experiment, used only aversive sound stimulus, 132 hybrid pigs (LD x LW), with 66 males and 66 females, were submitted to Y maze test to obtain a subgroup presenting divergent reactivity, considering the classification variables of reactivity defined in trial 1. Forty eight animals (24 females and 24 males) were identified as animals with high (HR) or low reactivity (LR) and submitted to two handling in pre-slaugther, a tranquil (T) and other stressful (S). There was no difference of reactivity between genders (male and female). HR and LR animals had similar (P> 0.05) for activity of plasma creatine kinase (1193.4 U / L ± 218 and 1539.9 U / L ± 218), lactate ( 3,91 mmol/L± 0,20 e 4,00 mmol/L± 0,21) and muscle glycogen 3 and 24 hours after slaughter (0,12 mg/g ± 0,02, 0,04 mg/g ± 0,01 e 0,08 mg/g±0,02, 0,04±0,01), respectively. Reactivity did not changed (P>0,05) the surface temperature of the animals (HR: 32,9 oC ± 0,46, LR: 32,9 oC ± 0,45), while the handling S resulted in surface temperature of 33,9 oC± 0,45, which exceeded (P<0,05) value of 31,9 o C ± 0,46 observed in the animals of handling T. Although tests have pointed to the possibility of categorizing reactivity by Y maze, the physiological measures examined were not influenced by divergent reactivity, even in pre-slaughter differentiated situations. The pre-slaugther handling did not influence the results of physiological variables, except for surface temperature, pointing to differences in the activity of the animals.

Analyzing the sentiment of Social Networks users is an attractive task, well-covered by the Sentiment Analysis research communities. Alongside, predicting the rating/opinion of users in Social Networks or e-commerce platforms is another attractive task covered by the Recommender Systems research communities. However, there is a rather new field of study that takes advantage of both of the mentioned scopes to predict the “unexpressed” opinion of users, based on their written sentiments and their similarity. Although the Social Network extracted data (due to the sparsity of the addressed items by different users) deals with high volumes of uncertainty, none of the few dozens of conducted studies in the Sentiment Prediction field focuses on managing the mentioned uncertainty. In this dissertation, we introduce the necessary foundations for constructing an Uncertainty-handling Sentiment Prediction system, by means of possibility theory, fuzzy theory, and probability theory. Moreover, we define an international project called probabilistic/possibilistic Text-based Emotion Rating (pTER) to fill and then enrich the gap of uncertainty management in Sentiment Prediction. pTER comprises two sub-projects: Scalar and Interval pTER. This dissertation provides five foundational research studies in the scalar pTER. Although the mentioned studies are sufficient for the targeted system, we let the scalar pTER system, itself, to be disseminated only after it can use its entire potency by utilizing the in-progress research projects of the other researchers of the pTER project, defined by this dissertation. In addition to the presented scalar-pTER studies, we also propose one research study in the interval pTER project which goes one step further in Uncertainty-handling and takes the measurement errors of the scalar pTER sub-systems into account. The presented studies in scalar- and interval-pTER belong to three phases: (I) Uncertainty-handling NLP platform, (II) Uncertainty-handling Sentiment Analysis, and (III) Uncertainty-handling Collaborative Filtering. The conducted experiments in this dissertation prove the superiority of our Uncertainty-handling approaches in all of these phases, in comparison to the corresponding state-of-the-art
Hay un campo de estudio bastante nuevo que aprovecha análisis de sentimiento y filtrado colaborativo para predecir la opinión "no expresada" de los usuarios, en función de sus sentimientos escritos y su similitud. Aunque de la red social se extraen datos (debido a la escasez de los elementos tratados por diferentes usuarios), abarcando e un alto volumen de incertidumbre, ninguna de las pocas docenas de estudios realizados en el campo predicción del sentimiento se centra en la gestión de la incertidumbre mencionada. Presentamos los fundamentos necesarios para construir un sistema de predicción del sentimiento de manejo de la incertidumbre, mediante la teoría de la posibilidad, la teoría difusa y la teoría de la probabilidad. Además, definimos un proyecto internacional llamado Probabilistic / Posibilista basado en el Emoción Rating (pTER) de textos para llenar y luego enriquecer el área de investigación de la gestión de la incertidumbre en predicción del sentimiento

CAPES - Coordenação de Aperfeiçoamento de Pessoal de Nível Superior
CNPq - Conselho Nacional de Desenvolvimento Científico e Tecnológico
FAPEMIG - Fundação de Amparo a Pesquisa do Estado de Minas Gerais
INCT-EIE - Instituto Nacional de Ciência e Tecnologia De Estruturas Inteligentes em Engenharia
Os sistemas de suspensão automotivos são projetados para permitir que o condutor tenha controle total do veículo. Esses sistemas devem garantir com eficiência o isolamento de vibração para o interior do automóvel, a aderência entre os pneus e o solo e a estabilidade do veículo durante a execução de manobras que envolvam mudança de direção. Neste trabalho foi conduzida uma investigação experimental sobre o comportamento dinâmico de um veículo de passeio da categoria hatch compacto premium. Manobras específicas de handling e ride foram consideradas nos testes físicos, a fim de avaliar a resposta do sistema de direção e o conforto dos passageiros. Handling se refere ao modo como o veículo se comporta durante a execução de manobras de mudança de direção e ride se relaciona à forma como a suspensão do veículo responde as irregularidades de pista durante o ato de dirigir. O trabalho de pesquisa contempla a definição de uma metodologia aplicada ao desenvolvimento de um modelo completo de veículo, através do método de multicorpos. Na modelagem matemática, a maioria dos componentes de suspensão foram considerados como rígidos, visto que as deformações elásticas são suficientemente pequenas, podendo assim serem consideradas desprezíveis. No entanto, é considerada a presença de alguns elementos flexíveis, em particular a barra estabilizadora e o eixo de torção, que se revelam essenciais na tentativa de construção de um modelo capaz de reproduzir adequadamente os efeitos não lineares. Nesse sentido, a barra estabilizadora foi representada como um corpo flexível, com o objetivo de incluir a sua não linearidade geométrica no modelo. Já a modelagem do eixo de torção foi baseada em uma técnica de cossimulação que utiliza um solver de multicorpos e um solver externo de elementos finitos para análise não linear. Simulações do comportamento dinâmico foram realizadas a fim de verificar a abrangência do modelo e correlação com os dados experimentais. O estudo de ride é apresentado enfatizando a métrica necessária para quantificar o conforto do veículo, de acordo com a norma ISO 2631-1. Os dados coletados experimentalmente foram processados no domínio da frequência e do tempo utilizando o software LMS Test.Lab. O processo de construção e simulação do modelo virtual foi realizado utilizando o software LMS Virtual.Lab Motion para a simulação de multicorpos e o LMS Samcef MECANO para a simulação de elementos finitos.
Automotive suspension systems are designed to allow the driver full control of the vehicle. These systems must efficiently ensure the vibration insulation of the car cabin, the tire grip on the road and the vehicle’s stability during turning maneuvers. In this study, an experimental investigation of the dynamic behavior of a subcompact premium car was conducted. Typical ride and handling maneuvers were considered in the physical tests in order to evaluate the steering response and passengers comfort. Vehicle handling refers to the way that the vehicle performs during turning maneuvers, and ride to how the vehicle suspension responds to the track irregularities during the act of driving. This work describes the methodology used for the development of a full vehicle model through the multibody method. In mathematical modeling, the majority of vehicle suspension components were included in the model as rigid bodies, due to the fact the elastic deformations are sufficiently small and hence negligible. However, there is the presence of some flexible elements, in particular the stabilizer bar and the twist beam, which are essential in the attempt to build a model capable of properly reproducing the nonlinear effects. Accordingly, the stabilizer bar was represented as a flexible body, in order to include geometric non-linearity into the model. The modeling of the twist beam was based on a co-simulation technique that uses a multibody solver and an external nonlinear solver for finite element method analysis. Dynamic behavior simulations were performed to verify the scope of the model and correlated with the experimental data. The ride comfort study was carried out emphasizing the required metric to quantify the vehicle comfort according to ISO 2631-1. The data collected were processed and analyzed in frequency and time domain using the LMS Test.Lab software. The building and simulation of the virtual model was performed using the LMS Virtual.Lab Motion software for the multibody simulation and LMS Samcef MECANO software for the finite element simulation.
Dissertação (Mestrado)

Submitted by Marco Antonio Oliveira da Silva (maosilva@ipen.br) on 2016-12-22T16:34:05Z No. of bitstreams: 0
Made available in DSpace on 2016-12-22T16:34:05Z (GMT). No. of bitstreams: 0
A reciclagem de refugos é um processo estratégico que viabiliza a expansão do mercado de plásticos, podendo gerar novos produtos. A obtenção de blendas poliméricas é uma alternativa nesse processo, já que é conhecida a perda de propriedades mecânicas dos termoplásticos nas etapas de reprocessamento. Neste trabalho, o polietileno de alta densidade (PEAD) reciclado teve adições de borracha do monômero etileno-propileno-dieno (EPDM) puro em baixos teores (1 %, 5 % e 10 %), cujo objetivo foi formar blendas miscíveis e um produto final mais resistente ao impacto. O PEAD foi submetido a quatro ciclos de moagem, extrusão e injeção (reprocessamento) e misturado ao EPDM puro por extrusão sem o uso de qualquer aditivo. Os grânulos da blenda formada foram usados para confeccionar os corpos-de-prova por injeção para seu uso nas distintas metodologias analíticas. O processo de irradiação gama foi aplicado nas doses de 50 kGy e 100 kGy e os parâmetros físico-químicos e mecânicos dessas amostras foram comparados aos das não irradiadas. Tanto as blendas irradiadas como as não irradiadas mostraramse visualmente e microscopicamente homogêneas, indicando a compatibilidade da mistura, que também é verificada por seu comportamento térmico. Os parâmetros mecânicos provenientes dos ensaios de tração e flexão, foram semelhantes nas amostras do termoplástico virgem e do reciclado; o processo de irradiação nas blendas gerou sua reticulação, sendo esta verificada não só pelo aumento dos valores desses parâmetros como também pelo aumento da fração gel. A resistência ao impacto aumentou cerca de duas vezes e meia nas amostras com maior teor de EPDM e cerca de 6 a 7 vezes nestas mesmas amostras irradiadas a 50 kGy e 100 kGy respectivamente. O novo material obtido tem fortes indicativos para a sua utilização na pequena e média indústria de plástico, uma vez que já com 1 % de EPDM apresentam melhores características mecânicas em relação ao termoplástico reciclado quatro vezes e essas características foram incrementadas após o processo de irradiação gama.
Dissertação (Mestrado em Tecnologia Nuclear)
IPEN/D
Instituto de Pesquisas Energéticas e Nucleares - IPEN-CNEN/SP

Diverses études épidémiologiques ont mis en évidence une relation étroite entre pollution urbaine au monoxyde de carbone (CO) et mortalité cardiovasculaire. Récemment il a été mis en évidence, chez le rat, qu'une exposition prolongée à ce polluant urbain avait pour conséquence le développement d'un phénotype cellulaire pathologique, pouvant influencer la vulnérabilité du coeur à un stress aigu. L'objectif de nos travaux était donc i) d'évaluer l'impact de la pollution au CO, sur la sensibilité du myocarde de rats au syndrome d'ischémie-reperfusion (IR) ; et ii) d'évaluer les effets potentiellement cardioprotecteurs d'un exercice pratiqué régulièrement à intensité modérée, sur le remodelage phénotypique cellulaire myocardique. Pour cela, 187 rats Wistar ont été séparés en 3 groupes : des rats contrôles, des rats exposés pendant 4 semaines au CO (30-100 ppm), et des rats entraînés en endurance avant d'être exposés au CO. La sensibilité à l'IR était évaluée par ischémie régionale réalisée sur modèle de coeur isolé perfusé de Langendorff. La fonction et les mouvements calciques de cardiomyocytes isolés était évalués en condition basale et consécutivement à un protocole d'anoxie-réoxygénation. Les résultats de ce travail confirment l'apparition d’un phénotype pathologique chez les rats exposés de façon prolongée au CO. Ce phénotype pathologique caractérisé dans notre travail par une altération de l’homéostasie calcique et du statut redox cellulaire ainsi qu'une expression tissulaire de iNOS apparait comme à l'origine de la plus grande vulnérabilité du coeur à un stress d’IR. Un autre résultat majeur de ce travail est qu’une stratégie de cardioprotection par un exercice d'intensité modérée pratiqué de manière régulière, permet de prévenir le remodelage pathologique cardiomyocytaire et ainsi l'augmentation de la sensibilité du myocarde à l'IR
Epidemiological studies suggested that carbon monoxide (CO) urban air pollution is mainly related to cardiovascular mortality. In addition, recent experimental studies have highlighted that CO exposure was responsible for the development of cardiomyocytes’ pathological remodeling, which can render the heart more vulnerable to acute stresses. Therefore, the aim of this experimental work was to i) evaluate the impact of prolonged exposure to simulated CO urban pollution on the sensitivity of the myocardium to IR ; and ii) evaluate potential cardioprotective effects of regular bouts of endurance training in this model. 187 Wistar rats were separated into 3 groups : control rats, CO rats exposed during 4 weeks to CO (30-100 ppm), and CO exercised rats. Myocardial sensibility to IR was evaluated with a regional ischemia performed on a Langendorff model of isolated heart. Moreover, the cardiomyocytes’ function and calcium handling were evaluated at basal conditions, following a protocol of cellular anoxia and reoxygenation. The results of this study confirm that chronic exposure to CO is responsible for cardiac phenotypic changes, which are characterized in this work by an imbalance in the cardiomyocytes’ oxidative status, an impairment of calcium handling and iNOS expression. These phenotypic changes were associated in this work with higher heart vulnerability to IR. Another major result of this study is that regular bouts of endurance training conducted prior to CO exposure prevented the pathological cardiac remodeling, consequently leading to higher heart vulnerability due to IR

碩士
國立中正大學
資訊工程研究所
100
Software testing is the main activity to ensure the quality of software. Our research team have developed a testing tool that applied a black-box testing technique to automatically generate test cases for Java methods with data of integer type. This thesis expends the tool to handle data of real type. The most difficult issue of software testing is test case generation. Test cases includes test input and expected output. Based on constraint logic programming, the expended tool can use a uniform method to generate test input and expect output for data of real type. This tool first uses specification reader to read class diagram and OCL and generate method constraint graphs. The test path generator then systematically enumerates the test paths on the method constraint graphs. Constraints on every test path are transformed into a constraint language programming predicate. The solution of this predicate is the corresponding test input and expected output for the test path. Type consists of three steps to ensure the correctness of test input and the precision of expected output. Finally, test class generator generates a test class from test cases.

碩士
國立中正大學
資訊工程研究所
100
Software testing is the main approach to ensuring the quality of software. The most critical issue in software testing is test case generation. We have developed a prototypical black-box method-level test case generator to generate test case automatically. This prototype uses UML class diagrams and Object Constraint Language to specify the behavior of methods and uses Constraint Logic Programming to generate test input and expected output as test cases. However, this prototype cannot handle function calls in specification. This paper extends this prototype by handling function calls. Our approach generates a Constraint Logic Programming predicates for OCL specification of each method. This predicate simulates the behavior of the method so that we can generate test cases which satisfy the constraints in the specification of both caller and callee. Therefore, this extension significantly extends the class of programs we can handle.

The aim of this study was to investigate whether daily handling in the form of tactile stimulation can ifluence behaviour of adult common voles (Microtus arvalis). The voles were tested in three types of behavioural experiments - Acoustic startle test, Open Field test (OFT), and Elevated Plus Maze test (EPM). No significant impact of handling on personality traits was detected.

碩士
國立臺北科技大學
資訊工程系
106
Bad smells in exception handling code may decrease the robustness of a program. Robusta is a static analysis tool which has been shown to be very useful in detecting exception handling bad smells in Java programs. The impact of a bad smell needs to be exposed so that we can know if it is a real bug. In this thesis, we propose a method for exposing the impact of a bad smell by using AspectJ, that is, to dynamically inject a code to make an exception be thrown at a desired place so that the effect of the exception to the program can be revealed. In this study, an empirical study has also been conducted by applying the proposed method to two open source software - JFreeChart and Tomighty. The result shows the proposed method can successfully expose the impact of exception handling bad smells detected by Robusta.

碩士
元智大學
工業工程與管理學系
107
The weapon system is fundamental in national defense and the weapon system requires a series of environmental tests before deployment. The purpose of environmental test is to ensure that the weapon system can operate normally in different environments throughout the life cycle. In other words, the design margin of the hardware can be realized by environmental qualification test. In the past, when the product failed after the environmental test, there are two ways to correct the failure; re-design and select different parts to replace the original one. It wastes not only a lot of manpower, but also time waste. However, the failure might be random one that is irrelevant to design or part selection. If repair is carried out immediately and re-design is skipped after product failure, many manhours and a investment will be saved. This study used tree diagrams from QC 7-method tools to analyze the development process of the project A. The discussion is divided into four parts. First of all, this study finds the weak point of the original process and modify it. Secondly, this study explores the number of subjects, the test sequence, and the re-verified process of the environmental qualification test. Thirdly, according to the U.S military standard, this study establishes a new process that should be followed when the product fails in the environmental test. The process is implemented for the project A and its benefit evaluated. Finally, the new process is included in the environmental specification, so that it can be followed by other similar project.

Durch das öffentliche und politische Interesse an der Tierhaltung und den daraus resultierenden Gesetzen werden sich die Haltungssysteme von Schweinen zukünftig ändern. Das Wohlbefinden der Tiere soll dadurch gesteigert werden. Dies kann z. B. durch die Gruppenhaltung der Tiere in verschiedenen Produktionsabschnitten erreicht werden. Neben den positiven Effekten für die Tiere führt das Halten von Schweinen in Gruppen aber zwangsläufig zu agonistischen Interaktionen innerhalb der Gruppe und damit zu Stress für die Tiere. Bei den wachsenden Bestandsgrößen ist ebenso eine gute Umgänglichkeit der Schweine wichtig, um Stress für Mensch und Tier beim Handling zu vermeiden. Die Zucht auf gruppentaugliche und umgängliche Schweine könnte dazu beitragen, die Probleme in diesen Bereichen zu reduzieren. Für eine Eingliederung von „Umgänglichkeit“ und „Aggressivität“ in Zuchtprogramme müssen zunächst Methoden gefunden werden, die zum Zielmerkmal korrelierte Merkmale erfassen, was bei einem vertretbaren Aufwand an einer großen Anzahl an Daten erfolgen kann. Außerdem müssen Zusammenhänge zu anderen ökonomisch wichtigen Zuchtmerkmalen analysiert und berücksichtigt werden. Ziel dieser Arbeit war es, Verhaltenstests zu entwickeln und deren Eignung als Selektionskriterium in Zuchtprogrammen zu überprüfen. Auf die Durchführung der Verhaltenstests unter praxisnahen Bedingungen und die Umsetzbarkeit zur Erfassung einer großen Anzahl an Verhaltensbeobachtungen wurde dabei besonders beachtet. Als erstes wurde der Einfluss des Haltungssystems auf das Verhalten der Tiere untersucht. Mit Hilfe von unterschiedlichen Verhaltenstest (Novel Object Test, Novel Arena Test, Wiegetest) sowie der Beurteilung der Anzahl und Schwere der Hautläsionen wurden 126 Mastschweine aus zwei unterschiedlich strukturierten Aufzuchtställen (10er Gruppen mit Trockenfutterautomaten vs. 30er Gruppen mit Trockenfutterautomaten und Breifutterautomaten) verglichen. Dadurch sollte aufgezeigt werden, in welchem Ausmaß das Verhalten der Tiere bereits durch kleine Veränderungen der Umwelt und der Gruppengröße beeinflusst wird. Während die Beobachtungen beim Novel Arena Test keine signifikanten Unterschiede aufwiesen, waren Schweine aus der angereicherten Umwelt deutlich aktiver z.B. beim Verhalten auf der Waage (p < 0,0001) und hatten auch weniger Hautläsionen (p = 0,0074) als die Tiere aus den 10er Gruppen (Kapitel 3). Des Weiteren wurde die Eignung von Hautläsionen als Indikator für individuelle Aggressivität überprüft. Dabei wurden Zusammenhänge zwischen dem agonistischen Verhalten beim Zusammenstallen von Sauen und deren Anzahl an Hautläsionen 10 Wochen nach der Gruppierung untersucht. Die Ergebnisse zeigten keinen Zusammenhang zwischen Anzahl der Initiierung von agonistischen Interaktionen bei der Gruppierung und der Anzahl an Hautläsionen beim Ausstallen aus der Gruppenhaltung (p > 0,1). Aber die häufiger attackierten Sauen hatten signifikant mehr Hautläsionen (p = 0,0435). Daher scheinen Anzahl und Schwere von Hautläsionen nicht als Merkmal für individuelle Aggressivität beim Schwein geeignet zu sein. Die aggressiveren Sauen hatten die besseren Fruchtbarkeitsleistungen (z.B. Häufigkeit der Initiierung von agonistischen Interaktionen vs. gesamt geborene Ferkel: rp = 0,20) und umgekehrt hatten die Sauen mit vielen Hautverletzungen schlechtere Fruchtbarkeitsleistungen (Score für Hautläsionen vorderes Körperdrittel vs. gesamt geborene Ferkel: rp = –0,28). Hier zeigte sich eine ungünstige Beziehung zwischen Aggressivität und Fruchtbarkeit (Kapitel 4). Im Weiteren wurden Zusammenhänge zwischen Backtest, Wiegeverhalten, Hautläsionen sowie Tageszunahmen analysiert. Dafür wurden sowohl genetische Parameter geschätzt als auch phänotypische Korrelationen berechnet. Verhaltensbeobachtungen von 976 Mastschweinen wurden über einen Zeitraum von der Geburt bis zum Schlachten der Tiere erfasst. Die geschätzten Heritabilitäten der Verhaltensmerkmale lagen in den meisten Fällen in einem züchterisch nutzbaren Bereich (z.B. Backtest 2: h² = 0,36 ± 0,08; Verhalten von Schlachtschweinen auf der Waage: h² = 0,20 ± 0,07). Die Ergebnisse zeigten aber, dass nur zwischen dem Backtest 1 und Backtest 2 (rp = 0,30; rg = 0,84 ± 0,11) sowie zwischen Backtest 2 und dem Verhalten von Läufern auf der Waage ein Zusammenhang besteht (rp = 0,13; rg = 0,57 ± 0,21). Aufgrund der vergleichsweise geringen Korrelationen zum Wiegenverhalten sowie den fehlenden Korrelationen zu weiteren Verhaltensmerkmalen, aber auch durch den hohen Zeit- und Arbeitsaufwand für die Durchführung, erscheint der Backtest insgesamt als Selektionsmerkmal ungeeignet. Weitere Zusammenhänge zwischen den verschiedenen Verhaltensmerkmalen konnten basierend auf den genetischen und phänotypischen Korrelationen nicht festgestellt werden. Die Tiere reagieren in den verschiedenen Situationen unterschiedlich, sodass kaum eine Konsistenz/Konstanz im Verhalten der Schweine bei unterschiedlichen Verhaltenstests beobachtet werden konnte. (Kapitel 5). Anhand der vorliegenden Ergebnisse, u.a. der Heritabilitäten, bleibt festzuhalten, dass die Integration von Verhaltensmerkmalen in Zuchtprogramme insgesamt möglich ist. Der erste wichtige Schritt ist die Auswahl eines Merkmals das in direkter Beziehung zu den Zielmerkmalen, Umgänglichkeit oder aggressiven Verhalten, steht. Sowohl der Backtest als auch die Hautläsionen sind nicht mit aggressiven Verhalten oder nur geringfügig mit Umgänglichkeit verbunden und sind somit nicht als Indikatormerkmale für eine Zucht auf Verhalten beim Schwein geeignet. Andere Merkmale, die z.B. beim Wiegetest oder bei der direkten Beobachtung des agonistischen Verhaltens erfasst werden, scheinen besser geeignet zu sein. Die positiven Korrelationen von Aggressivität zu Tageszunahmen und Fruchtbarkeit zeigen aber, dass ungünstige Beziehungen zwischen diesen ökonomisch wichtigen Merkmalen und ruhigem, friedlichem Verhalten bestehen.

Competitive and high-risk environments require complex high technology systems, which need to be supported and maintained over their respective life cycles. These systems often have a significant consequence of failure, and require complex management systems to achieve their operational objectives. Significant leadership and management challenges exist, not only in South Africa, but also in other developing countries, where systems may be utilised beyond the lifespan they were designed for and are susceptible to obsolescence. This study was conducted by following a structured process; the research consisted of three stages. The first stage dealt with the research problem, including the delimitations of the study. The second stage was further divided into three phases. The first phase deconstructed the appropriate literature, which included the interpretation of numerous definitions of logistics, integrated logistics support, and the integrated logistics support elements. In addition, the research was grounded in the fields of operations management, supply chain management and integrated logistics support. The second phase focused on the deconstruction of six case studies from four different high technology complex systems. From the analysis of the first two phases followed the third phase of research, which focused on the identification of areas requiring further research. Further research was conducted by means of a questionnaire, the results of which were analysed for variable dependency and variable association. The third stage of the research included the collation and analysis of the findings of the first two stages of research. The analysis utilised the principles of Mode 2 research and design science research, whereby an ILS framework and associated grounded technological rules have been recommended. These recommendations are robust in nature, as they can be applied in the most challenging environment and circumstances as identified. Furthermore, by grounding the theory in the disciplines of operations management, supply chain management and integrated logistics support, the reliability, validity, relevance, and applicability of the study could be substantiated. This implied that the generated theoretical knowledge could be transferred to and applied in practice, and as such, an organisation can reap substantial value added benefits, and gain considerable competitive advantage in the market place by applying this developed ILS framework and associated ILS grounded technological rules.
Business Leadership
D. BL.