No description available.

Dans ce mémoire, nous proposons une méthodologie statistique permettant d’obtenir un estimateur de l’espérance de vie des clients en assurance. Les prédictions effectuées tiennent compte des caractéristiques individuelles des clients, notamment du fait qu’ils peuvent détenir différents types de produits d’assurance (automobile, résidentielle ou les deux). Trois approches sont comparées. La première approche est le modèle de Markov simple, qui suppose à la fois l’homogénéité et la stationnarité des probabilités de transition. L’autre modèle – qui a été implémenté par deux approches, soit une approche directe et une approche par simulations – tient compte de l’hétérogénéité des probabilités de transition, ce qui permet d’effectuer des prédictions qui évoluent avec les caractéristiques des individus dans le temps. Les probabilités de transition de ce modèle sont estimées par des régressions logistiques multinomiales. / In this master’s thesis, we develop a statistical method to estimate the lifetime expectancy of clients in the insurance domain. The forecasts are personnalized according to the clients’ own features, the most notable being the fact that they can have any combination of automobile and residential insurance products. Three approaches are compared. The first approach is the simple Markov model which assume homogeneity and stationnarity of the transition probabilities. The other model suggested – which is implemented both by direct computation and by simulation – allows for heterogeneity of the transition probabilities, thus providing forecasts which evolve in time along with the characteristics of the clients. The transitions probabilities are estimated using multinomial logistic regressions.

Establishing group structure in complex networks is potentially very useful since nodes belonging to the same module can often be related by commonalities in their biological function. However, module detection in complex networks poses a challenging problem and has sparked a great deal of interest in various disciplines in recent years [5]. In real networks, which can be quite complex, we have no idea about the true number of modules that exist. Furthermore, the structure of the modules may be hierarchical meaning they may be further divided into sub-modules and so forth. Many attempts have been made to deal with these problems and because the involved methods vary considerably they have been difficult to compare [5]. The objectives of this thesis are (i) to create and implement a new algorithm that will identify modules in complex networks and reconstruct the network in such a way so as to maximize modularity, (ii) to evaluate the performance of a new method, and compare it to a popular method based on a simulated annealing algorithm, and (iii) to apply the new method, and a comparator method, to analyze the metabolic network of the bacterial genus Listeria, an important pathogen in both agricultural and human clinical settings.

No description available.

Šis darbas skirtas Italijos verslininkų pasitenkinimo darbu analizei, naudojantis kokybinių požymių daugiamatės statistikos metodais. Duomenys yra iš Italijos Veneto regiono verslininkų tyrimo, kurį 2006 metais pradėjo Padujos universiteto Statistikos departamentas. Imtį sudaro 1216 stebėjimų iš beveik 113000 Veneto regiono (Šiaurės rytų Italija) verslininkų populiacijos. Priklausomasis kintamasis, t.y. pasitenkinimas darbu, išmatuotas ranginėje keturių taškų Likert skalėje su šiomis kategorijomis: „nepatenkintas“, „nei patenkintas, nei nepatenkintas“, „pakankamai patenkintas“ ir „patenkintas“. Paaiškinamuosius kintamuosius sudaro demografinės, firmos charakteristikos, tapimo verslininku motyvus, darbą, laisvalaikį ir ateities perspektyvas atspindintys kintamieji. Beveik visi kintamieji, taip pat ir priklausomasis, yra kategoriniai. Todėl pagrindinis šio darbo tikslas yra išsirinkti geriausiai tokio tipo duomenims tinkantį modelį. Aptariama galimybė naudotis tiesine regresija, atlikus optimalųjį matavimo skalės transformavimą (optimal scaling), logistine regresija, prieš tai apjungus priklausomojo kintamojo rangus, daugianariu logit modeliu, analizuojant kiekvienos kategorijos tikimybę atskirai ir rangine regresija su logit ir cloglog sąryšio funkcijomis. Potencialių pasitenkinimo darbu veiksnių skaičius- pakankamai didelis (daugiau negu 30), todėl itin svarbus uždavinys yra parinkti aiškinančiuosius kintamuosius, vengiant multikolinearumo problemos. Siekiant sumažinti... [toliau žr. visą tekstą] / The paper presents job satisfaction analysis among Italian entrepreneurs using multivariate statistical techniques. Data are taken from Veneto region (North-East part of Italy) businessmen research started by Department of Statistics of Padova University at 2006. Sample consists of 1216 observations (real population is almost 113 000 entrepreneurs at this region). The outcome variable for the job satisfaction is measured on an ordered, categorical, four-point Likert scale – „dissatisfied‘“, „neither satisfied nor dissatisfied“ , „quite satisfied“ and „satisfied“. Explanatory variables include demographic items, firm characteristics, variables representing the reasons of having started the own business, items associated with work, leisure activities and future perspectives. Quite all variables, as well as the dependent one, beeing categorical, the main objective of this work is to select an appropriate model for such type of data among possible alternatives. Discussions are made on possibility to applicate linear regression with optimal scaling, binary logit for dichotomized dependent variable, multinomial logit for analysis of every single category and ordinal logistic regression with several link functions (logit and cloglog). The number of possible determinants of job satisfaction beeing quite large (there were more than 30 questionnaire items associated with job satisfaction) it was of a great importance choosing which explanatory variables should be included in the models... [to full text]

This thesis investigates how a tax and warning label on less healthy snack food products may affect consumer behaviour when the imposition of the tax is a source of consumer information. A survey that included choice experiments was implemented in supermarkets. Participants were asked to choose between high fat snacks, some displaying a stigmatizing warning label, and healthier snacks. Multinomial logit and latent class models exploring choice were estimated and a predictive hypothetical market was set up. Results show that the warning label had a negative price premium of about $4. The effect of price, though small, becomes even smaller as BMI increases. A fat tax for health is not recommended because it might not hit the target population, people were not very price sensitive, and it would likely be regressive. To encourage health, it appears to be more effective to display a warning label than to apply a tax. / Agricultural and Resource Economics

New Zealand is internationally renowned for having some of the finest and most challenging trout fishing in the world. However, due to continuing development and angling pressure many fishing sites are showing signs of environmental degradation and over fishing. This trend is almost certain to continue into the future given continued population and economic growth. Understanding the determinants of site choice, preference heterogeneity and anglers’ substitution patterns is fundamentally important to fishery managers who have the difficult task of maintaining quality angling experiences on a number of fishing sites, managing angling pressure and maintaining license sales. Recent advances in simulation techniques and computational power have improved the capability of discrete choice models to reveal preference heterogeneity and complex substitution patterns among individuals. This thesis applies and evaluates a number of state-of-the-art discrete choice models to study angler site choice in New Zealand. Recreation specialisation theory is integrated into the analysis to enhance the behavioural representation of the statistical models. A suite of models is presented throughout the empirical portion of this thesis. These models demonstrate different ways and degrees of explaining preference heterogeneity as well as identifying anglers’ substitution patterns. The results show that North Canterbury anglers’ preferences vary considerably. Resource disturbances such as riparian margin erosion, reduced water visibility and declines in catch rates can cause significant declines in angler use of affected sites, and at the same time non-proportional increases in the use of unaffected sites. Recreation specialisation is found to be closely related to the types of fishing site conditions, experiences and regulations preferred by anglers. Anglers’ preference intensities for fishing site attributes, such as catch rates, vary across different types of fishing sites. This location specific preference heterogeneity is found to be related to specialisation. Overall, the empirical findings indicate that conventional approaches to modelling angler site choice which do not incorporate a strong understanding of angler preference heterogeneity can lead to poorly representative models and suboptimal management and policy outcomes.

Nas últimas décadas inúmeros estudos foram produzidos acerca da influência do local de moradia sobre o bem-estar das pessoas, o chamado “efeito-bairro”. Este trabalho tem o objetivo de analisar os principais fatores que influenciam os jovens gaúchos de 15 a 24 anos a permanecerem apenas estudando, estudando e trabalhando, só trabalhando ou não estudando e nem trabalhando. Serão analisados fatores ligados ao local de moradia, características individuais dos jovens, tais como raça, sexo e idade, e familiares, tais como educação dos pais e renda familiar per capta e como cada uma dessas variáveis exerce influência sobre as escolhas dos jovens das aglomerações urbanas do Rio Grande do Sul. Para tanto, serão utilizados os microdados amostrais dos Censos Demográficos de 2000 e 2010 do IBGE, através dos quais serão construídas as variáveis dependente e independentes que constituirão a análise estatística via modelos de regressão logística multinomial. Os resultados das influências dos fatores ligados ao local de moradia apontaram que residir longe do centro, em 2000, dificultava o ingresso no mercado de trabalho para o jovem de Porto Alegre, enquanto em 2010 aumenta suas chances de estudar e de trabalhar e quanto mais elevado é o nível socioeconômico da vizinhança do jovem, maiores são suas chances de apenas estudar e menores suas chances de só trabalhar. O estudo também constatou que o jovem ser filho ou enteado do chefe domiciliar aumenta significativamente suas chances de estudar e reduz bastante suas chances de dedicarem-se exclusivamente ao trabalho; quanto mais elevada a idade do jovem, maiores suas chances de trabalhar e menores de só estudar, consequência da transição natural do jovem da escola para o mercado de trabalho; o jovem ser do sexo masculino aumenta suas chances de trabalhar; quanto mais elevada a renda domiciliar per capita do jovem, menores são suas chances de não estudar nem trabalhar; quanto mais anos de estudo o chefe domiciliar possuir, maiores as chances do jovem estudar; quanto mais crianças no domicílio do jovem, maiores são suas chances de não trabalhar nem estudar e quanto mais elevado o grupo de categoria sócio-ocupacional do chefe domiciliar, maiores são as chances do jovem estudar. / In recent decades numerous studies have been made about the influence of the place of residence on the well-being of people, the so-called "neighborhood effect." This work aims to analyze the main factors influencing the gauchos youth 15 to 24 years to stay just studying, studying and working, just working or no studying and neither working. We will analyze factors related to place of residence, individual characteristics of young people, such as race, gender and age, and family, such as parental education and family income per capita and how each of these variables influences the choices of young agglomerations urban of Rio Grande do Sul. Therefore, the sample microdata from Demographic Censuses of 2000 and 2010 IBGE will be used, through which the dependent and independent variables that constitute the statistical analysis via multinomial logistic regression models will be built. The results of the influence of factors related to place of residence indicated that reside far from the center, in 2000, made it difficult to enter the labor market for young Porto Alegre, while in 2010 increases your chances of study and work and the higher It is the socioeconomic status of the neighborhood of the young, the greater your chances of just studying and lower your chances of just work. The study also found that young to be a child or stepchild of the household head significantly increases your chances of study and greatly reduces your chances to devote themselves exclusively to the work; the higher the age of the young, the greater your chances of work and under only studying result of the natural transition of the young from school to the labor market; the young being male increases your chances of working; the higher the household income per capita of the young, the lower your chances of not study or work; the more years of schooling the household head has, the more likely the young study; the more children in the household of the young, the greater your chances of not working or studying and the higher the socio-occupational category group head home, the greater the chances of the young study.

Submitted by Daniel Felipe Silva Santos null (danielfssantos@yahoo.com.br) on 2017-08-29T19:56:20Z No. of bitstreams: 1 ReconhecedorDeVeiculos2D-DBM.pdf: 3800862 bytes, checksum: 46f12ff55f4e0680833b9b1b184ad505 (MD5) / Approved for entry into archive by Luiz Galeffi (luizgaleffi@gmail.com) on 2017-08-29T20:19:13Z (GMT) No. of bitstreams: 1 santos_dfs_me_sjrp.pdf: 3800862 bytes, checksum: 46f12ff55f4e0680833b9b1b184ad505 (MD5) / Made available in DSpace on 2017-08-29T20:19:13Z (GMT). No. of bitstreams: 1 santos_dfs_me_sjrp.pdf: 3800862 bytes, checksum: 46f12ff55f4e0680833b9b1b184ad505 (MD5) Previous issue date: 2017-08-14 / Neste trabalho é proposto um método para reconhecer veículos em imagens coloridas baseado em uma rede neural Perceptron Multicamadas pré-treinada por meio de técnicas de aprendizado em profundidade, sendo uma das técnicas composta por Máquinas de Boltzmann Profundas e projeção bilinear e a outra composta por Máquinas de Boltzmann Profundas Multinomiais e projeção bilinear. A proposição deste método justifica-se pela demanda cada vez maior da área de Sistemas de Transporte Inteligentes. Para se obter um reconhecedor de veículos robusto, a proposta é utilizar o método de treinamento inferencial não-supervisionado Divergência por Contraste em conjunto com o método inferencial Campos Intermediários, para treinar múltiplas instâncias das redes profundas. Na fase de pré-treinamento local do método proposto são utilizadas projeções bilineares para reduzir o número de nós nas camadas da rede. A junção das estruturas em redes profundas treinadas separadamente forma a arquitetura final da rede neural, que passa por uma etapa de pré- treinamento global por Campos Intermediários. Na última etapa de treinamentos a rede neural Perceptron Multicamadas (MLP) é inicializada com os parâmetros pré-treinados globalmente e a partir deste ponto, inicia-se um processo de treinamento supervisionado utilizando gradiente conjugado de segunda ordem. O método proposto foi avaliado sobre a base BIT-Vehicle de imagens frontais de veículos coletadas de um ambiente de tráfego real. Os melhores resultados obtidos pelo método proposto utilizando rede profunda multinomial foram de 81, 83% de acurácia média na versão aumentada da base original e 91, 10% na versão aumentada da base combinada (Carros, Caminhões e Ônibus). Para a abordagem de redes profundas não multinomiais os melhores resultados foram de 81, 42% na versão aumentada da base original e 91, 13% na versão aumentada da base combinada. Com a aplicação da projeção bilinear, houve um decréscimo considerável nos tempos de treinamento das redes profundas multinomial e não multinomial, sendo que no melhor caso o tempo de execução do método proposto foi 5, 5 vezes menor em comparação com os tempos das redes profundas sem aplicação de projeção bilinear. / In this work it is proposed a vehicle recognition method for color images based on a Multilayer Perceptron neural network pre-trained through deep learning techniques (one technique composed by Deep Boltzmann Machines and bilinear projections and the other composed by Multinomial Deep Boltzmann Machines and bilinear projections). This proposition is justified by the increasing demand in Traffic Engineering area for the class of Intelligent Transportation Systems. In order to create a robust vehicle recognizer, the proposal is to use the inferential unsupervised training method of Contrastive Divergence together with the Mean Field inferential method, for training multiple instances of deep models. In the local pre-training phase of the proposed method, bilinear projections are used to reduce the number of nodes of the neural network. The combination of the separated trained deep models constitutes the final recognizer’s architecture, that yet will be global pre-trained through Mean Field. In the last phase of training the Multilayer Perceptron neural network is initialized with globally pre-trained parameters and from this point, a process of supervised training starts using second order conjugate gradient. The proposed method was evaluated over the BIT-Vehicle database of frontal images of vehicles collected from a real road traffic environment. The best results obtained by the proposed method that used multinomial deep models were 81.83% of mean accuracy in the augmented original database version and 91.10% in the augmented combined database version (Cars, Trucks and Buses). For the non-multinomial deep models approach, the best results were 81.42% in the augmented version of the original database and 91.13% in the augmented version of the combined database. It was also observed a significant decreasing in the training times of the multinomial deep models and non-multinomial deep models with bilinear projection application, where in the best case scenario the execution time of the proposed method was 5.5 times lower than the deep models that did not use bilinear projection.

Nas últimas décadas inúmeros estudos foram produzidos acerca da influência do local de moradia sobre o bem-estar das pessoas, o chamado “efeito-bairro”. Este trabalho tem o objetivo de analisar os principais fatores que influenciam os jovens gaúchos de 15 a 24 anos a permanecerem apenas estudando, estudando e trabalhando, só trabalhando ou não estudando e nem trabalhando. Serão analisados fatores ligados ao local de moradia, características individuais dos jovens, tais como raça, sexo e idade, e familiares, tais como educação dos pais e renda familiar per capta e como cada uma dessas variáveis exerce influência sobre as escolhas dos jovens das aglomerações urbanas do Rio Grande do Sul. Para tanto, serão utilizados os microdados amostrais dos Censos Demográficos de 2000 e 2010 do IBGE, através dos quais serão construídas as variáveis dependente e independentes que constituirão a análise estatística via modelos de regressão logística multinomial. Os resultados das influências dos fatores ligados ao local de moradia apontaram que residir longe do centro, em 2000, dificultava o ingresso no mercado de trabalho para o jovem de Porto Alegre, enquanto em 2010 aumenta suas chances de estudar e de trabalhar e quanto mais elevado é o nível socioeconômico da vizinhança do jovem, maiores são suas chances de apenas estudar e menores suas chances de só trabalhar. O estudo também constatou que o jovem ser filho ou enteado do chefe domiciliar aumenta significativamente suas chances de estudar e reduz bastante suas chances de dedicarem-se exclusivamente ao trabalho; quanto mais elevada a idade do jovem, maiores suas chances de trabalhar e menores de só estudar, consequência da transição natural do jovem da escola para o mercado de trabalho; o jovem ser do sexo masculino aumenta suas chances de trabalhar; quanto mais elevada a renda domiciliar per capita do jovem, menores são suas chances de não estudar nem trabalhar; quanto mais anos de estudo o chefe domiciliar possuir, maiores as chances do jovem estudar; quanto mais crianças no domicílio do jovem, maiores são suas chances de não trabalhar nem estudar e quanto mais elevado o grupo de categoria sócio-ocupacional do chefe domiciliar, maiores são as chances do jovem estudar. / In recent decades numerous studies have been made about the influence of the place of residence on the well-being of people, the so-called "neighborhood effect." This work aims to analyze the main factors influencing the gauchos youth 15 to 24 years to stay just studying, studying and working, just working or no studying and neither working. We will analyze factors related to place of residence, individual characteristics of young people, such as race, gender and age, and family, such as parental education and family income per capita and how each of these variables influences the choices of young agglomerations urban of Rio Grande do Sul. Therefore, the sample microdata from Demographic Censuses of 2000 and 2010 IBGE will be used, through which the dependent and independent variables that constitute the statistical analysis via multinomial logistic regression models will be built. The results of the influence of factors related to place of residence indicated that reside far from the center, in 2000, made it difficult to enter the labor market for young Porto Alegre, while in 2010 increases your chances of study and work and the higher It is the socioeconomic status of the neighborhood of the young, the greater your chances of just studying and lower your chances of just work. The study also found that young to be a child or stepchild of the household head significantly increases your chances of study and greatly reduces your chances to devote themselves exclusively to the work; the higher the age of the young, the greater your chances of work and under only studying result of the natural transition of the young from school to the labor market; the young being male increases your chances of working; the higher the household income per capita of the young, the lower your chances of not study or work; the more years of schooling the household head has, the more likely the young study; the more children in the household of the young, the greater your chances of not working or studying and the higher the socio-occupational category group head home, the greater the chances of the young study.