本文探討大專生的學業成就，如何受到學生本身的外表吸引力以及工讀的型態所影響。本研究使用國立台灣師範大學台灣高等教育資料庫所建構的「九十二學年度大三學生問卷調查」以及「九十三學年度大專畢業生畢業後一年問卷調查」，以univariate probit model進行分析。研究結果發現：一、具外表吸引力的學生，在平均成績超過80分的機率，比外表較不具外表吸引力的學生高出5.6%；二、具外表吸引力的學生，在進入班級排名前10%的機率較不具外表吸引力的學生高出3.7%；三、雖然從事校外工讀對於學業成就有負面影響，不過從事校內工讀反而對於學業成就有正面影響。 接著以內生性分析的結果顯示，外表吸引力以及校外工讀與學業成就之間存在內生性問題。在recursive bivariate probit model估計下，外表吸引力與學業成就之間的關係仍舊為正向顯著，且男學生的外表吸引力對於學業成就之影響高於女學生；至於校外工讀與學業成就之間的關係在recursive bivariate probit model估計下反而轉變為正，表示在考慮內生性問題之後，校外工讀反而對於學業成就具有正向影響。 最後以trivariate probit model探討外表吸引力、校外工讀以及學業成就三者之間的關係，結果發現在控制外表吸引力下，校外工讀對於學業成就的影響差異並不大，工讀的負面效果並不會因為具有外表吸引力而有所減緩。

Ocean wave energy is considered to be one of the important potential renewable energy resources for sustainable development. Various wave energy converter technologies have been proposed to harvest the energy from ocean waves. This thesis is based on the linear generator wave energy converter developed at Uppsala University. The research in this thesis focuses on the foundation optimization and the power absorption optimization of the wave energy converters and on the wave climate modelling at the Lysekil wave converter test site. The foundation optimization study of the gravity-based foundation of the linear wave energy converter is based on statistical analysis of wave climate data measured at the Lysekil test site. The 25 years return extreme significant wave height and its associated mean zero-crossing period are chosen as the maximum wave for the maximum heave and surge forces evaluation. The power absorption optimization study on the linear generator wave energy converter is based on the wave climate at the Lysekil test site. A frequency-domain simplified numerical model is used with the power take-off damping coefficient chosen as the control parameter for optimizing the power absorption. The results show a large improvement with an optimized power take-off damping coefficient adjusted to the characteristics of the wave climate at the test site. The wave climate modelling studies are based on the wave climate data measured at the Lysekil test site. A new mixed distribution method is proposed for modelling the significant wave height. This method gives impressive goodness of fit with the measured wave data. A copula method is applied to the bivariate joint distribution of the significant wave height and the wave period. The results show an excellent goodness of fit for the Gumbel model. The general applicability of the proposed mixed-distribution method and the copula method are illustrated with wave climate data from four other sites. The results confirm the good performance of the mixed-distribution and the Gumbel copula model for the modelling of significant wave height and bivariate wave climate.

本文以1997年6月至2016年7月臺灣上櫃股票市場做為研究樣本，透過建立一Bivariate Diagonal BEKK GARCH (1,1)-in-mean模型，並以大盤週轉率形成之總合流動性指標與大盤超額報酬率之共變異數做為系統流動性風險之衡量指標，觀察系統流動性風險在臺灣上櫃股票市場是否有被訂價。結論除發現系統流動性風險有確實被訂價外，系統流動性風險溢價還兼具穩定性，且對市場超額報酬率有顯著的影響力。 / By constructing a bivariate diagonal BEKK Garch (1,1)-in-mean model and using the covariance between the excess market return and turnover rate as aggregate systematic liquidity proxy, the study tries to examine whether systematic liquidity risk was priced on Taiwan OTC stock market during the period of June 1997-July 2016. Based on monthly data, the findings suggest that not only the systematic liquidity risk was well priced on Taiwan OTC stock market, but the phenomenon also possessed stability and could have significant impact on stock returns.

Esta tese trata de dois assuntos importantes na teoria de risco: o fenômeno da dependência local e a identificação e mensuração de assimetrias apresentadas pelos dados. A primeira parte trata de dependência local, sendo abordadas algumas medidas já analisadas na literatura. Versões locais dos coeficientes de Kendall e Spearman , baseadas na distribuição condicional dos dados, são propostas. São apresentadas algumas propriedades dessas medidas e a aplicação das mesmas a algumas cópulas. Na segunda parte são apresentados resultados sobre cópulas bivariadas que são as menos associativas e menos bi-simétricas segundo o critério de máxima distância modular. A última parte trata da não-permutabilidade e assimetria radial dos dados. Uma medida de não-permutabilidade baseada nos coeficientes de correlação condicional é proposta e aplicada a algumas distribuições. No final, o conceito de quantil bivariado é aplicado nas definições de medidas para avaliar o grau de permutabilidade e de simetria radial presentes na estrutura de dependência dos dados e de testes de hipóteses para verificar se a cópula subjacente aos dados é permutável ou radialmente simétrica. / In this thesis two important fields in risk theory are studied: the local dependence phenomenon and the identification and measuring of asymmetries contained in data. The first part deals with local dependence: some measures already studied in the literature are presented and discussed, and local versions of the coefficients Kendall and Spearman , based on the conditional distribution of data, are proposed. Properties of these measures and some examples concerning its application are treated. In the second part are presented some results about bivariate copulas which are the least associative and the least bi-symmetric according to the maximum modular distance. The last part analyses the nonexchangeability and the radial asymmetry of data. A measure of nonexchangeability based on the conditional correlation coefficient is proposed and applied to some distribution functions. At the end, the concept of bivariate quantile is applied in the definitions of measures for evaluating the degree of exchangeability and radial symmetry present in data and of hypothesis tests proposed for verifying whether the underlying copula is exchangeable or radially symmetric.

Neste trabalho são reunidos diferentes modelos e técnicas para representar situações experimentais ou observacionais de análise de sobrevivência. Para modelar respostas bivariadas e covariáveis foi proposto o modelo de regressão Kumaraswamy-Weibull bivariado. A presen»ca de indivíduos curados foi considerada sob duas diferentes abordagens, originando o modelo de regressão com fração de cura para dados bivariados por meio de cópulas e o modelo de regressão log-linear bivariado com fração de cura. Os parâmetros dos modelos foram esti- mados pelo método de máxima verossimilhança sujeito a restriçãoo nos parâmetros por meio da função barreira adaptada. Adaptou-se uma análise de sensibilidade de forma a considerar as metodologias de Influência Global, Influência Local e Influência Local Total para verificar vários aspectos que envolvem a formulação e ajuste dos modelos propostos. Utilizou-se um conjunto de dados de insuficiência renal e retinopatia diabética são utilizados para exemplificar a aplicação dos modelos propostos. / This work brought together di®erent models and techniques to represent expe- rimental or observational situations in survival analysis. To model bivariate responses and covariates was proposed Kumaraswamy Weibull bivariate regression model. The presence of cured individuals was considered under two di®erent approaches originating the regression model with a cured fraction for bivariate data through copulas and the log-linear bivariate regression model with cured fraction. The parameters of the models were estimated by ma- ximum likelihood method subject to the restriction on the parameters through the adapted barrier function. A sensitivity analysis was adapted considering the methodologies of Global In°uence, Local In°uence and Total Local In°uence to check various aspects of the formulation and adjustment of the models proposed. Data set of renal failure and diabetic retinopathy are used to exemplify the application of the proposed models.

Made available in DSpace on 2017-05-12T14:47:07Z (GMT). No. of bitstreams: 1 protegidoJacqueline_dissertacao.pdf: 3012964 bytes, checksum: 337f371fa8c665bb4fdfdc22709938da (MD5) Previous issue date: 2015-02-02 / The soil spatial variability s studies are based in geostatistics which appears as a method whose data comes from natural phenomena and consider the geographical location of the phenomenon. If in an area under study the researcher has interest in searching the variability of variables and has evidence that the steps which describe the spatial structure of this variables aren t independent a geostatistical bivariate model study can be proposed. This work concentrates on evaluating the variation s influence of the bivariate Gaussian common component model (BGCCM) parameters in calculating the Pearson correlation coefficient and analyzing the influence that sizes and sampling settings may present at the BGCCM s estimation and at spatial prediction variables in non-sampled locations. Moreover, for co-placed samples, crossed semivariograms were built and compared with univariate model and BGCCM, in relation to estimates of the model and the sizes associated with the spatial prediction. In order to do it, these methodologies were applied in simulated data sets and experimental data, from an agricultural property. The simulation study of the parameters variation influence s analysis of the bivariate model BGCCM in calculating the Pearson correlation coefficient between described variables of bivariate model BGCCM revealed that the Pearson s linear correlation coefficient can t be considered in decision-making about the presence of joint spatial dependence between pairs of variables. In the study with simulations, it was observed that the biggest differences of accuracy measures and the square sum of the spatial prediction s difference occurred when the univariate models and crossed semivariogram were compared to the BGCCM. Moreover, the simulation s study observed that for balanced data the regular and irregular meshes showed better efficiency as the spatial prediction. The study with real data showed that under the BGCCM approach, spatial dependence was observed, mainly between pH and Mn for co-placed and balanced data of the agriculture year 2010/2011; and between the variable inside of the next pairs: (Prod, RSP0-10), (Prod, RSP11-20), (Prod, RSP21-30) and (Prod, Mn) for co-placed and balanced data of the agriculture year 2013/2014; and (Prod, RSP11-20) and (Prod, RSP21-30) for co-placed and unbalanced data of the agriculture year 2013/2014. Still considering the real data study comparing the univariated models, crossed semivariogram and BGCCM, differences could be observed in the square sum of prediction s difference and in the accuracy measures, both for balanced and unbalanced data. However, considering the real data and the control sample, the spatial prediction s quality using the BGCCM model was inefficient when compared to the quality resulted from the spatial prediction using the univariate model. Nevertheless, this result may have been influenced by the choice of sample configuration. / Estudos da variabilidade espacial do solo estão baseados na geoestatística, que se apresenta como um método cujos dados provêm de fenômenos naturais e que consideram a localização geográfica do fenômeno. Se numa área em estudo o pesquisador tem interesse em pesquisar a variabilidade das variáveis e se há evidências que os passos que descrevem a estrutura espacial dessas variáveis não são independentes pode-se propor o estudo de um modelo geoestatístico bivariado. Este trabalho concentrou-se em avaliar a influência da variação dos parâmetros do modelo bivariado com componente de correlação parcialmente comum (bivariate Gaussian common component model BGCCM) no cálculo do coeficiente de correlação linear de Pearson e analisar a influência que tamanhos e configurações amostrais podem apresentar na estimação do modelo BGCCM e na predição espacial de variáveis em localizações não amostradas. Além disso, para amostras co-locadas, construíram-se os semivariogramas cruzados e comparou-se com o modelo univariado e BGCCM, em relação às estimativas do modelo e as medidas associadas à predição espacial. Para isso, essas metodologias foram aplicadas em conjuntos de dados simulados e dados experimentais, provenientes de uma propriedade agrícola. O estudo de simulação da análise da influência da variação dos parâmetros do modelo bivariado BGCCM no cálculo do coeficiente de correlação linear de Pearson entre as variáveis descritas do modelo bivariado BGCCM revelou que o coeficiente de correlação linear de Pearson não pode ser considerado na tomada de decisão quanto à presença de dependência espacial conjunta entre pares de variáveis. No estudo com simulações observou-se que as maiores diferenças das medidas de acurácia e da soma quadrada da diferença entre as predições espaciais ocorreram quando se comparou os modelos univariado e semivariograma cruzado com o BGCCM. Ainda no estudo de simulação observou-se que para os dados balanceados as malhas regular e irregular apresentaram melhor eficiência quanto à predição espacial. O estudo com dados reais mostrou que, sob a abordagem do modelo BGCCM, observou-se a presença de dependência espacial principalmente entre pH e Mn para dados co-locados e balanceados do ano agrícola 2010/2011; e entre as variáveis dentro dos seguintes pares: (Prod, RSP0-10), (Prod, RSP11-20), (Prod, RSP21-30) e (Prod, Mn) para dados co-locados e balanceados do ano agrícola 2013/2014; e (Prod, RSP11-20) e (Prod, RSP21-30) para dados co-locados e desbalanceados do ano agrícola 2013/2014. Ainda considerando o estudo com dados reais comparando os modelos univariado, semivariograma cruzado e BGCCM, mostraram diferenças na soma quadrada da diferença da predição e nas medidas acurácia, tanto para dados balanceados como para os desbalanceados. No entanto, considerando os dados reais e a amostra controle, a qualidade da predição espacial usando o modelo BGCCM se mostrou ineficiente quando comparada com a qualidade obtida na predição espacial usando o modelo univariado. Porém, esse resultado pode ter sido influenciado pela escolha da configuração amostral utilizada.

One basic problem in statistical sciences is to understand the relationships among multivariate outcomes. Although it remains an important tool and is widely applicable, the regression analysis is limited by the basic setup that requires to identify one dimension of the outcomes as the primary measure of interest (the "dependent" variable) and other dimensions as supporting this variable (the "explanatory" variables). There are situations where this relationship is not of primary interest. For example, in actuarial sciences, one might be interested to see the dependence between annual claim numbers of a policyholder and its impact on the premium or the dependence between the claim amounts and the expenses related to them. In such cases the normality hypothesis fails, thus Pearson's correlation or concepts based on linearity are no longer the best ones to be used. Therefore, in order to quantify the dependence between non-normal outcomes one needs different statistical tools, such as, for example, the dependence concepts and the copulas. This thesis is devoted to modelling dependence with applications in actuarial sciences and is divided in two parts: the first one concerns dependence in frequency credibility models and the second one dependence between continuous outcomes. In each part of the thesis we resort to different tools, the stochastic orderings (which arise from the dependence concepts), and copulas, respectively. During the last decade of the 20th century, the world of insurance was confronted with important developments of the a posteriori tarification, especially in the field of credibility. This was dued to the easing of insurance markets in the European Union, which gave rise to an advanced segmentation. The first important contribution is due to Dionne & Vanasse (1989), who proposed a credibility model which integrates a priori and a posteriori information on an individual basis. These authors introduced a regression component in the Poisson counting model in order to use all available information in the estimation of accident frequency. The unexplained heterogeneity was then modeled by the introduction of a latent variable representing the influence of hidden policy characteristics. The vast majority of the papers appeared in the actuarial literature considered time-independent (or static) heterogeneous models. Noticeable exceptions include the pioneering papers by Gerber & Jones (1975), Sundt (1988) and Pinquet, Guillén & Bolancé (2001, 2003). The allowance for an unknown underlying random parameter that develops over time is justified since unobservable factors influencing the driving abilities are not constant. One might consider either shocks (induced by events like divorces or nervous breakdown, for instance) or continuous modifications (e.g. due to learning effect). In the first part we study the recently introduced models in the frequency credibility theory, which can be seen as models of time series for count data, adapted to actuarial problems. More precisely we will examine the kind of dependence induced among annual claim numbers by the introduction of random effects taking unexplained heterogeneity, when these random effects are static and time-dependent. We will also make precise the effect of reporting claims on the a posteriori distribution of the random effect. This will be done by establishing some stochastic monotonicity property of the a posteriori distribution with respect to the claims history. We end this part by considering different models for the random effects and computing the a posteriori corrections of the premiums on basis of a real data set from a Spanish insurance company. Whereas dependence concepts are very useful to describe the relationship between multivariate outcomes, in practice (think for instance to the computation of reinsurance premiums) one need some statistical tool easy to implement, which incorporates the structure of the data. Such tool is the copula, which allows the construction of multivariate distributions for given marginals. Because copulas characterize the dependence structure of random vectors once the effect of the marginals has been factored out, identifying and fitting a copula to data is not an easy task. In practice, it is often preferable to restrict the search of an appropriate copula to some reasonable family, like the archimedean one. Then, it is extremely useful to have simple graphical procedures to select the best fitting model among some competing alternatives for the data at hand. In the second part of the thesis we propose a new nonparametric estimator for the generator, that takes into account the particularity of the data, namely censoring and truncation. This nonparametric estimation then serves as a benchmark to select an appropriate parametric archimedean copula. This selection procedure will be illustrated on a real data set.

This dissertation consists of four papers, each with an application of a discrete dependent variable model, censored regression or duration model to a credit market phenomenon or monetary policy question. The first three essays deal with bank lending policy, while the last one studies interest rate policy by Central Banks. In the first essay, a bivariate probit model is estimated to contrast the factors that influence banks’ loan granting decision and individuals’ risk of default. This model is used as a tool to construct a Value at Risk measure of the credit risk involved in a portfolio of consumer loans and to investigate the efficiency of bank lending policy. The second essay takes the conclusions from the first paper as a starting point. It investigates if the fact that banks do not minimize default risk can be explained by the existence of return maximization policy. For this purpose, a Tobit model with sample selection effects and variable censoring limits is developed and estimated on the survival times of consumer loans. The third paper focuses on dormancy, instead of default risk or survival time, as the most important factor affecting risk and return in bank lending. By means of a duration model the factors determining the transition from an active status to dormancy are studied. The estimated model is used to predict the expected durations to dormancy and to analyze the expected profitability for a sample loan applicants. In the fourth paper, the discrete nature of Central Bank interest rate policy is studied. A grouped data model, that can take the long periods of time without changes in the repo rate by the Central Bank into account, is estimated on weekly Swedish data. The model is found to be reasonably good at predicting interest rate changes. / Diss. (sammanfattning) Stockholm : Handelshögsk.

Η παρούσα διπλωματική εργασία αντλεί την θεματολογία της από την θεωρία ελέγχων καλής προσαρμογής. Δίνονται τα βασικά σημεία της θεωρίας ελεγχοσυναρτήσεων και στη συνέχεια παρουσιάζεται η επέκταση του έλεγχου των Kolmogorov-Smirnov στο διδιάστατο χώρο καθώς και μια τροποποίησή της. Βασικό βοήθημα για την επέκταση αυτή αποτελεί το θεώρημα του Rosenblatt, το οποίο προτείνει ένα μετασχηματισμό μιας απόλυτα συνεχούς k-διάστατης κατανομής σε ομοιόμορφη κατανομή στον k-διάστατο υπερκύβο. Παρουσιάζεται επίσης το στατιστικό Α, το οποίο προτάθηκε από τον Damico. Η ιδιαιτερότητα αυτού του στατιστικού είναι ότι έχει διακριτή κατανομή. Προτείνεται ένα στατιστικό για τον έλεγχο καλής προσαρμογής συνεχών δεδομένων αρχικά στις δύο και στη συνέχεια στις k διαστάσεις. Ως εργαλεία χρησιμοποιήθηκαν το στατιστικό Α και το Θεώρημα του Rosenblatt. Για διάφορα μεγέθη δείγματος, δίνονται ο πίνακας πιθανοτήτων για τις τιμές του στατιστικού καθώς και ο πίνακας με τις κρίσιμες τιμές για διάφορες τιμές του p-value. Οι πίνακες αυτοί προέκυψαν κυρίως με μεθόδους προσομοίωσης. Τέλος, υπολογίστηκε η ισχύς του ελέγχου και γίνεται σύγκριση με την ισχύ του διδιάστατου Kolmogorov-Smirnov. / This project is based in theory of goodness-fit-tests. We present the most important componenets of test funcion theory. Also, we present the extension of the Kolmogorov-Smirnov test in bivariate case and an approximation. This extension is based on Rosenblatt's theorem, which suggests a transformation of an absolutly continious k-variate distribution into the uniform distribution of the k-dimentional hypercube. Moreover, is presented the statistic A, which was suggested from Damico. The particularity of this statistic is that has a district contribution. We suggest a goodnes-of-fit test for continious data first on two dimensions and after on k dimensions. This new statistic uses Rosenblatt's transformation and the statistic A. For different sizes of sample, are given the table of probablities and the table with the critical values. These tables were arised with simulation methods. Finally, was computed the power of the test and was compared with the power of the bivariate Kolmogorv-Smirnov.

Mελετάται η μίξη δύο συνεχών κατανομών με αύξοντα ρυθμό αποτυχίας και δίνονται συνθήκες για να έχει η μίξη φθίνοντα ρυθμό αποτυχίας. Όταν η μία από τις δύο κατανομές της μίξης είναι η εκθετική γίνεται αντιστροφή του ρυθμού αποτυχίας. Στην περίπτωση της μίξης δύο κανονικών κατανομών παρουσιάζεται ο τρόπος που συνδέεται το πλήθος των κορυφών της πυκνότητας με τον ρυθμό αποτυχίας της μίξης. Mελετάται επίσης, η μονοτονία του ρυθμού αποτυχίας διακριτών κατανομών χρησιμοποιώντας τον λόγο δύο διαδοχικών πιθανοτήτων και δίδεται μία συνθήκη για να έχει η μίξη δύο διακριτών κατανομών φθίνοντα ρυθμό αποτυχίας όταν η μία από τις δύο κατανομές της μίξης είναι η γεωμετρική. Τέλος, χρησιμοποιώντας τον λόγο διαδοχικών πιθανοτήτων, μελετούμε την μονοτονία του ρυθμού αποτυχίας για διδιάστατες διακριτές κατανομές. / The mixture of two continuous distributions, with increasing failure rates, is considered and the necessary conditions to have decreasing failure rate (DFR) are given. When one of these distributions is the Exponential, reversal of the failure rate is observed. In the case of two normal distributions the failure rate is associated with the number of modes. It is also considered the failure rate for discrete distributions in regard with the ratio of two consecutive probabilities. A condition to have DFR is given when one of the distributions of the mixture is the geometric. Finally, we make use of the ratio of two consecutive probabilities to study the failure rate for bivariate discrete distributions.