|
|
Parametric risk modelling in assessing mortality
Hlavandová, Radana ; Mazurová, Lucie (advisor) ; Branda, Martin (referee)
In this thesis we focus on modeling stochastic mortality and parameter risk in assessing mortality. We explore two mortality stochastic models for modeling the number of deaths in portfolio which consist of one or more than one cohort. We define the term mixture of distributions and introduce Beta-Binomial and Poisson-Gamma model. We address immediate life annuities and we apply Bayesian Poisson- Gamma model to quantify longevity risk on data. The obvious increasing trend of average lifetime leads insurance companies to greater protection against longevity risk. We show how to deal with solvency rules by internal models designed consistently with the requirement in the standard formula of Solvency II. Powered by TCPDF (www.tcpdf.org)
|
|
|
Study of the dependence structure in economic and financial data
Hlavandová, Radana ; Zichová, Jitka (advisor) ; Petrásek, Jakub (referee)
Title: Study of the dependence structure in economic and financial data Author: Radana Hlavandová Department: Department of Probability and Mathematical Statistics Supervisor: RNDr. Jitka Zichová, Dr., Department of Probability and Mathematical Statistics Abstract: The thesis focuses on the issue of graphical models as a possible \\method for determining relationships between different variables. The thesis provides a broad theoretical basis for two methods of testing data, the test of zero partial correlation coefficients and the test based on maximum likelihood estimate. The last mentioned approach is a test of a graphical model with a data set on the basis of deviance. The thesis describes the theory of conditional independence and Markov properties as the basis of both tests, which are illustrated by general examples and by an example with real financial data. Keywords: partial correlation coefficients, conditional independence graph, graphical models
|
|
|
Parametric risk modelling in assessing mortality
Hlavandová, Radana ; Mazurová, Lucie (advisor) ; Branda, Martin (referee)
In this thesis we focus on modeling stochastic mortality and parameter risk in assessing mortality. We explore two mortality stochastic models for modeling the number of deaths in portfolio which consist of one or more than one cohort. We define the term mixture of distributions and introduce Beta-Binomial and Poisson-Gamma model. We address immediate life annuities and we apply Bayesian Poisson- Gamma model to quantify longevity risk on data. The obvious increasing trend of average lifetime leads insurance companies to greater protection against longevity risk. We show how to deal with solvency rules by internal models designed consistently with the requirement in the standard formula of Solvency II. Powered by TCPDF (www.tcpdf.org)
|
|
|
Study of the dependence structure in economic and financial data
Hlavandová, Radana ; Zichová, Jitka (advisor) ; Petrásek, Jakub (referee)
Title: Study of the dependence structure in economic and financial data Author: Radana Hlavandová Department: Department of Probability and Mathematical Statistics Supervisor: RNDr. Jitka Zichová, Dr., Department of Probability and Mathematical Statistics Abstract: The thesis focuses on the issue of graphical models as a possible \\method for determining relationships between different variables. The thesis provides a broad theoretical basis for two methods of testing data, the test of zero partial correlation coefficients and the test based on maximum likelihood estimate. The last mentioned approach is a test of a graphical model with a data set on the basis of deviance. The thesis describes the theory of conditional independence and Markov properties as the basis of both tests, which are illustrated by general examples and by an example with real financial data. Keywords: partial correlation coefficients, conditional independence graph, graphical models
|
guest ::
