Na literatura da área da análise de sobrevivência existem os modelos tradicionais, ou sem fração de cura, e os modelos de longa duração, ou com fração de cura. Recentemente tem sido proposto um modelo mais geral, conhecido como o modelo com fatores de risco latentes com esquemas de ativação. Nesta tese são deduzidas novas propriedades que possuem a função de sobrevivência, a função de taxa de risco e o valor esperado, quando e considerado o modelo com fatores de risco latentes. Estas propriedades são importantes, já que muitos outros modelos que tem aparecido na literatura recentemente podem ser considerados como casos particulares do modelo com fatores de risco latentes. Além disto, são propostos novos modelos de sobrevivência e estes são aplicados a conjuntos de dados reais. Também é realizado um estudo de simulação e uma análise de sensibilidade, para mostrar a qualidade destes modelos / In the survival literature we can find traditional models without cure fraction and longterm models with cure fraction. A more general risk factor model with latent activation scheme has been recently proposed. In this thesis we deduce new properties for the survival function, hazard function and expected value for this model. Since many recent survival models can be regarded as particular cases of the risk factor model with latent activation scheme these properties are of great relevance. In addition we propose new survival models that are applied to real data examples. A simulation and sensibility analysis are also performed to asses the goodness of fit of these models

No presente trabalho, e proposto um modelo de regressão utilizando a distribuição gama generalizada exponenciada (GGE) para dados censurados, esta nova distribuição e uma extensão da distribuição gama generalizada. A distribuição GGE (CORDEIRO et al., 2009) que tem quatro parâmetros pode modelar dados de sobrevivência quando a função de risco tem forma crescente, decrescente, forma de U e unimodal. Neste trabalho apresenta-se uma expansão natural da distribuição GGE para dados censurados, esta distribuição desperta o interesse pelo fato de representar uma família paramétrica que possui como casos particulares outras distribuições amplamente utilizadas na analise de dados de tempo de vida, como as distribuições gama generalizada (STACY, 1962), Weibull, Weibull exponenciada (MUDHOLKAR et al., 1995, 1996), exponencial exponenciada (GUPTA; KUNDU, 1999, 2001), Rayleigh generalizada (KUNDU; RAKAB, 2005), dentre outras, e mostra-se útil na discriminação entre alguns modelos probabilísticos alternativos. Considerando dados censurados, e abordado o método de máxima verossimilhança para estimar os parâmetros do modelo proposto. Outra proposta deste trabalho e introduzir um modelo de regressão log-gama generalizado exponenciado com efeito aleatório. Por fim, são apresentadas três aplicações para ilustrar a distribuição proposta. / In the present study, we propose a regression model using the exponentiated generalized gama (EGG) distribution for censored data, this new distribution is an extension of the generalized gama distribution. The EGG distribution (CORDEIRO et al., 2009) that has four parameters it can model survival data when the risk function is increasing, decreasing, form of U and unimodal-shaped. In this work comes to a natural expansion of the EGG distribution for censored data, is awake distribution the interest for the fact of representing a parametric family that has, as particular cases, other distributions which are broadly used in lifetime data analysis, as the generalized gama (STACY, 1962), Weibull, exponentiated Weibull (MUDHOLKAR et al., 1995, 1996), exponentiated exponential (GUPTA; KUNDU, 1999, 2001), generalized Rayleigh (KUNDU; RAKAB, 2005), among others, and it is shown useful in the discrimination among some models alternative probabilistics. Considering censored data, the maximum likelihood estimator is considered for the proposed model parameters. Another proposal of this work was to introduce a log-exponentiated generalized gamma regression model with random eect. Finally, three applications were presented to illustrate the proposed distribution.

Αρχικά γίνεται μια ιστορική αναδρομή, μια παρουσίαση της διπαραμετρικής Γενικευμένης εκθετικής κατανομής (τύπος κατανομής, συνάρτηση πυκνότητας πιθανότητας κλπ) και αναφέρονται βασικά χαρακτηριστικά της κατανομής. Στη συνέχεια αναφέρονται βασικοί ορισμοί και θεωρήματα σχετικά κυρίως με τη σημειακή παραμετρική εκτίμηση καθώς και την εκτίμηση κατά Bayes. Το επόμενο κεφάλαιο πραγματεύεται την ανάλυση του μοντέλου και τις βασικές ιδιότητες της Γενικευμένης εκθετικής κατανομής. Επίσης μελετώνται ειδικά θέματα, όπως συναρτήσεις επιβίωσης, πληροφορία Fisher, διατεταγμένες παρατηρήσεις, κατανομή του αθροίσματος και παραγωγή τυχαίων αριθμών, στα πλαίσια της Γενικευμένης εκθετικής κατανομής. Στη συνέχεια αναλύονται και εφαρμόζονται μέθοδοι σημειακής εκτίμησης (Μέγιστη Πιθανοφάνεια, Μέθοδος ροπών, Μέθοδος εκατοστημορίων, Ελάχιστα και σταθμισμένα ελάχιστα Τετράγωνα, L-ροπές) για την εκτίμηση των παραμέτρων της κατανομής. Μελετάται και η απόδοση των εκτιμητών για τις διάφορες μεθόδους εκτίμησης. Ακολουθεί η εκτίμηση τύπου Bayes των παραμέτρων (με συναρτήσεις ζημίας τετραγωνικού σφάλματος και LINEX αντίστοιχα). Αναφέρονται πάλι συμπεράσματα για την απόδοση των εκτιμητών και σύγκριση με τους εκτιμητές μέγιστης πιθανοφάνειας. Τελικά παρουσιάζουμε την προσέγγιση ενός αναλογιστικού πίνακα μέσω της Γενικευμένης εκθετικής κατανομής. / In the beginning, we mention a historical recursion, a presentation of the 2-parameter Generalized exponential distribution ( distribution type, probability density function etc.) and we also mention basic characteristics of the distribution. Basic definitions and theorems about point estimation and Bayes estimation are reported. Furthermore, we discource on the analysis of the model and basic properties of the Generalized exponential distribution. Special themes, such as survival functions, Fisher information, order statistics, sum distribution and production of random numbers are analyzed in the frame of the Generalized exponential distribution. Moreover, we analyze and apply point estimation methods (maximum likelihood, method of moments, percentile estimation, least (and weighted least) squares, method of L-moments) in order to estimate parameters of the distribution. Performance of the estimators for different estimation methods is also analyzed. Next, bayesian estimation of the parameters (under squared error loss function and LINEX loss function) is coming up for discussion. We also analyze the performance of the estimators and compare them to the maximum likelihood estimators. Finally, we present approximation of an actuarial table via Generalized exponential distribution.

Na literatura da área da análise de sobrevivência existem os modelos tradicionais, ou sem fração de cura, e os modelos de longa duração, ou com fração de cura. Recentemente tem sido proposto um modelo mais geral, conhecido como o modelo com fatores de risco latentes com esquemas de ativação. Nesta tese são deduzidas novas propriedades que possuem a função de sobrevivência, a função de taxa de risco e o valor esperado, quando e considerado o modelo com fatores de risco latentes. Estas propriedades são importantes, já que muitos outros modelos que tem aparecido na literatura recentemente podem ser considerados como casos particulares do modelo com fatores de risco latentes. Além disto, são propostos novos modelos de sobrevivência e estes são aplicados a conjuntos de dados reais. Também é realizado um estudo de simulação e uma análise de sensibilidade, para mostrar a qualidade destes modelos / In the survival literature we can find traditional models without cure fraction and longterm models with cure fraction. A more general risk factor model with latent activation scheme has been recently proposed. In this thesis we deduce new properties for the survival function, hazard function and expected value for this model. Since many recent survival models can be regarded as particular cases of the risk factor model with latent activation scheme these properties are of great relevance. In addition we propose new survival models that are applied to real data examples. A simulation and sensibility analysis are also performed to asses the goodness of fit of these models

No presente trabalho, e proposto um modelo de regressão utilizando a distribuição gama generalizada exponenciada (GGE) para dados censurados, esta nova distribuição e uma extensão da distribuição gama generalizada. A distribuição GGE (CORDEIRO et al., 2009) que tem quatro parâmetros pode modelar dados de sobrevivência quando a função de risco tem forma crescente, decrescente, forma de U e unimodal. Neste trabalho apresenta-se uma expansão natural da distribuição GGE para dados censurados, esta distribuição desperta o interesse pelo fato de representar uma família paramétrica que possui como casos particulares outras distribuições amplamente utilizadas na analise de dados de tempo de vida, como as distribuições gama generalizada (STACY, 1962), Weibull, Weibull exponenciada (MUDHOLKAR et al., 1995, 1996), exponencial exponenciada (GUPTA; KUNDU, 1999, 2001), Rayleigh generalizada (KUNDU; RAKAB, 2005), dentre outras, e mostra-se útil na discriminação entre alguns modelos probabilísticos alternativos. Considerando dados censurados, e abordado o método de máxima verossimilhança para estimar os parâmetros do modelo proposto. Outra proposta deste trabalho e introduzir um modelo de regressão log-gama generalizado exponenciado com efeito aleatório. Por fim, são apresentadas três aplicações para ilustrar a distribuição proposta. / In the present study, we propose a regression model using the exponentiated generalized gama (EGG) distribution for censored data, this new distribution is an extension of the generalized gama distribution. The EGG distribution (CORDEIRO et al., 2009) that has four parameters it can model survival data when the risk function is increasing, decreasing, form of U and unimodal-shaped. In this work comes to a natural expansion of the EGG distribution for censored data, is awake distribution the interest for the fact of representing a parametric family that has, as particular cases, other distributions which are broadly used in lifetime data analysis, as the generalized gama (STACY, 1962), Weibull, exponentiated Weibull (MUDHOLKAR et al., 1995, 1996), exponentiated exponential (GUPTA; KUNDU, 1999, 2001), generalized Rayleigh (KUNDU; RAKAB, 2005), among others, and it is shown useful in the discrimination among some models alternative probabilistics. Considering censored data, the maximum likelihood estimator is considered for the proposed model parameters. Another proposal of this work was to introduce a log-exponentiated generalized gamma regression model with random eect. Finally, three applications were presented to illustrate the proposed distribution.

En el presente documento se emplean modelos de supervivencia para analizar la duración en que la tasa de interés de referencia, utilizada por el Banco Central de Reserva del Perú como instrumento de política monetaria, permanece sin cambios. Para el análisis se emplean modelos no paramétricos y paramétricos, permitiendo la naturaleza de datos censurados por la derecha y covariables no constantes en el tiempo. Para el análisis de supervivencia del modelo no paramétrico, se emplea el estimador Kaplan-Meier para formar las funciones de supervivencia y de riesgo. Por otro lado, en cuanto al análisis de los modelos paramétricos, se comparan dichas funciones estimadas bajo una función Exponencial, Weibull y Log-logística. Para ello, se emplea como covariables la variación mensual y anual del producto bruto interno (PBI), la inflación, la tasa de interés de referencia, el desempleo y el tipo de cambio. Se estiman 24 modelos y se selecciona el mejor de acuerdo con la significancia de las variables y el criterio de información de Akaike. Se obtiene que tanto para el análisis no paramétrico y paramétrico, la probabilidad de que la tasa de interés de referencia permanezca sin cambios es cada vez menor a lo largo del tiempo. Además, en el modelo paramétrico bajo la distribución Weibull y Loglogística (distribución escogida como preferida) se obtienen como variables significativas la inflación, el producto bruto interno y el nivel de la tasa de interés de referencia; sin embargo, al emplear la distribución Exponencial, el producto bruto interno no es significativo. / In the present document, survival models are used to analyze the duration of the reference interest rate, while it remains constant, used by the Central Reserve Bank of Perú as its monetary policy instrument. For the analysis, both nonparametric and parametric models are estimated, allowing the nature for right-censoring of the data and time-varying covariates. In case of non-parametric model, Kaplan-Meier estimator is used to model survival and hazard functions. In case of parametric models, the survival and hazard functions are compared under an Exponential, Weibull and Log-Logistical functions. The monthly and annual variation of the gross domestic product, the inflation rate, the interest rate, the unemployment rate and the exchange rate are used as covariates. Twenty-four models are estimated. The best one is selected according to the significance of the covariate’s ant Akaike information criterion. The results show that for both non-parametric and parametric models, the probability of a constant interest rate remains unchanged is less over the time. Furthermore, in the parametric model under Weibull distribution and Log-Logistical distribution (preferred distribution), inflation rate, gross domestic product and the interest rate are obtained as significant variables; however, the gross domestic product isn´t significant under Exponential distribution. / Trabajo de investigación

No description available.

In this thesis, we develop some inferential results for one-shot device testing data analysis. These extend and generalize existing methods in the literature. First, a competing-risk model is introduced for one-shot testing data under accelerated life-tests. One-shot devices are products which will be destroyed immediately after use. Therefore, we can observe only a binary status as data, success or failure, of such products instead of its lifetime. Many one-shot devices contain multiple components and failure of any one of them will lead to the failure of the device. Failed devices are inspected to identify the specific cause of failure. Since the exact lifetime is not observed, EM algorithm becomes a natural tool to obtain the maximum likelihood estimates of the model parameters. Here, we develop the EM algorithm for competing exponential and Weibull cases. Second, a semi-parametric approach is developed for simple one-shot device testing data. Semi-parametric estimation is a model that consists of parametric and non-parametric components. For this purpose, we only assume the hazards at different stress levels are proportional to each other, but no distributional assumption is made on the lifetimes. This provides a greater flexibility in model fitting and enables us to examine the relationship between the reliability of devices and the stress factors. Third, Bayesian inference is developed for one-shot device testing data under exponential distribution and Weibull distribution with non-constant shape parameters for competing risks. Bayesian framework provides statistical inference from another perspective. It assumes the model parameters to be random and then improves the inference by incorporating expert's experience as prior information. This method is shown to be very useful if we have limited failure observation wherein the maximum likelihood estimator may not exist. The thesis proceeds as follows. In Chapter 2, we assume the one-shot devices to have two components with lifetimes having exponential distributions with multiple stress factors. We then develop an EM algorithm for developing likelihood inference for the model parameters as well as some useful reliability characteristics. In Chapter 3, we generalize to the situation when lifetimes follow a Weibull distribution with non-constant shape parameters. In Chapter 4, we propose a semi-parametric model for simple one-shot device test data based on proportional hazards model and develop associated inferential results. In Chapter 5, we consider the competing risk model with exponential lifetimes and develop inference by adopting the Bayesian approach. In Chapter 6, we generalize these results on Bayesian inference to the situation when the lifetimes have a Weibull distribution. Finally, we provide some concluding remarks and indicate some future research directions in Chapter 7. / Thesis / Doctor of Philosophy (PhD)

In the case of soil chemical analyses, frequency distributions for some elements show a dissymmetrical aspect, with a very marked spread to the right or to the left. A high frequency of extreme values is also observed and a possible mixture of several distributions, due to the presence of various soil types within a single geographical unit, is encountered. Then, for the outliers detection and the establishment of detection limits, an original outliers detection procedure has been developed; it allows estimating extreme quantiles above and under which observations are considered as outliers. The estimation of these detection limits is based on the right and the left of the distribution tails. A first estimation is realised for each elementary geographical unit to determine an appropriate truncation level. Then, a spatial classification allows creating adjoining homogeneous groups of geographical units to estimate robust limit values based on an optimal number of observations. / Dans le cas des analyses chimiques de sols, les distributions de fréquences des résultats présentent, pour certains éléments étudiés, un caractère très dissymétrique avec un étalement très marqué à droite ou à gauche. Une fréquence importante de valeurs extrêmes est également observée et un mélange éventuel de plusieurs distributions au sein dune même entité géographique, lié à la présence de divers types de sols, peut être rencontré. Dès lors, pour la détection des valeurs aberrantes et la fixation des limites de détection, une méthode originale, permettant destimer des quantiles extrêmes au-dessus et en dessous desquelles les observations sont considérées comme aberrantes, a été élaborée. Lestimation des limites de détection est établie de manière distincte à partir des queues des distributions droite et gauche. Une première estimation par entité géographique élémentaire est réalisée afin de déterminer un niveau de troncature adéquat. Une classification spatiale permet ensuite de créer des groupes dentités homogènes contiguës, de manière à estimer des valeurs limites robustes basées sur un nombre dobservations optimal.

Made available in DSpace on 2016-06-02T20:04:51Z (GMT). No. of bitstreams: 1 3028.pdf: 7339557 bytes, checksum: 16711c2271b754604bfa0b0fba30290b (MD5) Previous issue date: 2010-03-04 / In this work, we introduce a Bayesian Analysis for survival multivariate data in the presence of a covariate vector and censored observations. Different frailties or latent variables are considered to capture the correlation among the survival times for the same individual. We also introduce a Bayesian analysis for some of the most popular bivariate exponential distributions introduced in the literature. A Bayesian analysis is also introduced for the Block & Basu bivariate exponential distribution using Markov Chain Monte Carlo (MCMC) methods and considering lifetimes in presence of covariates and censored data. In another topic, we introduce a Bayesian Analysis for bivariate lifetime data in the presence of covariates and censoring data assuming different bivariate Weibull distributions derived from some existing copula functions. A great computational simplification to simulate samples for the joint posterior distribution is obtained using the WinBUGS software. Numerical illustrations are introduced considering real data sets considering every proposed methodology. / Nesta tese introduzimos uma an´alise Bayesiana para dados de sobreviv encia multivariados, na presen¸ca de um vetor de covari´aveis e observa¸c oes censuradas. Diferentes fragilidades ou vari´aveis latentes s ao consideradas para capturar a correla¸c ao existente entre os tempos de sobreviv encia, para o mesmo indiv´ıduo. Tamb´em apresentamos uma an´alise Bayesiana para algumas das mais populares distribui¸c oes exponenciais bivariadas introduzidas na literatura. Uma an´alise Bayesiana tamb´em ´e introduzida para a distribui¸c ao exponencial bivariada de Block & Basu, usando m´etodos MCMC (Monte Carlo em Cadeias de Markov) e considerando os tempos de sobreviv encia na presen¸ca de covari´aveis e dados censurados. Em outro t´opico, introduzimos uma an´alise Bayesiana para dados de sobreviv encia bivariados na presen¸ca de covari´aveis e observa¸c oes censuradas, assumindo diferentes distribui¸c oes bivariadas Weibull derivadas de algumas fun¸c oes c´opulas existentes. Uma grande simplifica¸c ao computacional para simular amostras da distribui¸c ao a posteriori conjunta de interesse ´e obtida usando o software WinBUGS. Ilustra¸c oes num´ericas s ao introduzidas considerando conjunto de dados reais, para cada uma das metodologias propostas.