|
|
Maximum likelihood estimators and their approximations
Tyuleneva, Anastasia ; Omelčenko, Vadim (advisor) ; Zvára, Karel (referee)
Title: Maximum likelihood estimators and their approximations Author: Anastasia Tyuleneva Department: Department of Probability and Mathematical Statistics Supervisor: Mgr. Vadym Omelchenko Abstract: Maximum likelihood estimators method is one of the most effective and accurate methods that was used for estimation distributions and parameters. In this work we will find out the pros and cons of this method and will compare it with other estimation models. In the theoretical part we will review important theorems and definitions for creating common solution algorithms and for processing the real data. In the practical part we will use the MLE on the case study distributions for estimating the unknown parameters. In the final part we will apply this method on the real price data of EEX A. G, Germani. Also we will compare this method with other typical methods of estimation distributions and parameters and chose the best distribution. All tests and estimators will be provided by Mathematica software. Keywords: parametr estimates, Maximum Likelihood estimators, MLE, Stable distribution, Characteristic function, Pearson's chi-squared test, Rao-Crámer. .
|
|
|
Backtesting of Different Scaling Rules for Value at Risk in the Basel Context
Klečka, Adam ; Krištoufek, Ladislav (advisor) ; Avdulaj, Krenar (referee)
1 Abstract There is a discrepancy between two important horizon for Value at Risk modelling in the Basel context. We take 10-day values for determining the regulatory capital but we consider 1-day models for backtesting. The main objective of this thesis is to examine the suitability of the currently used Square Root of Time rule for Value at Risk scaling. We compare its performance with the method utilizing Hurst exponent. Our analysis is performed for both the normal and stable distribution. We conclude that the normality assumption and the Square Root of Time rule prevail under the regulatory parameters. The results of the Hurst exponent method are not favourable under normality. On the other hand, the performance for the stable distribution is quite satisfactory under non-Basel parameters and the Hurst exponent complements this distribution very well. Therefore, the use of stable distribution and the Hurst exponent method is justified when dealing with complex non-linear instruments, during turbulent periods, or for general non-Basel setting. In general however, our results are strongly data-dependent and further evidence is needed for any conclusive implications. JEL Classification G21, G28, C58, G32, C14, G18 Keywords value at risk, backtesting, volatility scaling, Basel II, stable distribution, Hurst...
|
|
|
Stable distributions for feature extraction from speech signals
Mžourek, Z.
The aim of this paper is to introduce class of stable distributions as a potentional tool for statistical modelling of features extracted from speech signals. Alpha-stable distributions are generalization of the Gaussian distribution therefore they can be used in modeling of more variety of different problems. It is described why can stable distributions be useful in speech processing and potential useful applications are proposed for feature extractions and reduction.
|
|
|
Elliptical Stable Distributions
Omelchenko, Vadym
The elliptical stable distributions represent a symmetric subfamily of the stable distributions. Their advantage contrary to the general stable distributions consists in their easy-to-use property and the highest resemblance to the normal distribution. They enable an easy representation of the dependence structure of the margins by means of a matrix Q the same as in case of the normal distribution. In general, the dependence structure between margins is given in form of a spectral measure which can be even continuous. The computations and approximations require so much time that it just the fact that many practitioners avoid using general stable distributions. The general stable distributions possess so many additional properties that they barely take after the multivariate normal distribution. But the multi-variate elliptical stable distributions can be easily simulated and the estimation of their parameters can be obtained by methods whose preciseness is almost the same as the one of the maximum likelihood methodology.
|
|
|
Empirical Estimates in Economic and Financial Problems via Heavy Tails
Kaňková, Vlasta
Optimization problems depending on a probability measure correspond to many economic and financial applications. Complete knowledge of this measure is necessary to solve exactly these problems. Since this condition is fulfilled only seldom, the problem has to be usually solved on the data basis to obtain satistical estimates of an optimal value and optimal solutions. Great effort has been paid to investigate properties of these estimates; first under assumptions of disribution with thin tails and linear dependence on the probability measure. Recently, it has appeared an investigation in the case of nonlinear dependence on the probability measure and heavy tailed distributions with shape parameter greater two. We focus on the case of the stable and Pareto distributions with a shape parameter in the inteval (1, 2).
|
|
| |
|
|
Stable distributions: On densities
Karlová, Andrea
William Feller stated that densitites of stable distributions are probably not known. V.M. Zolotarev spent a lot of eort on deriving aproximative formulas via series expansions and integral representations for stable densities. In this report we investigate derivation of stable densities and stable distribution function. We further investigate relation between stable distribution functions and incomplete gamma functions.
|
guest ::
