|
|
Statistical inference for random processes
Kvitkovičová, Andrea ; Štěpán, Josef (referee) ; Hlubinka, Daniel (advisor)
The thesis deals with testing hypotheses about the parameters of the Wiener process with a constant drift rate and instantaneous variance. The tests are based on the first time, when the process reaches a pre-specified boundary point. We consider a process with a non-negative drift rate, and we observe hitting a positive point. We focus on tests about the drift rate, in particular about the absence of any drift. We first study several basic properties of the Wiener process and its connection with the Wiener process with a drift. Using these, we derive distributional properties of the first hitting time. We also describe selected hypotheses testing techniques in the setting of exponential families. We construct uniformly most powerful unbiased tests of one parameter in the presence of a nuisance parameter. Further, we construct uniformly most powerful tests of hypotheses about the drift rate, while the variance is known, and we study this situation in more detail. Finally, we construct asymptotic simultaneous tests of both parameters based on the R'enyi divergences.
|
|
|
Periodic regression quantiles
Kotík, Lukáš ; Jurečková, Jana (referee) ; Hlubinka, Daniel (advisor)
The thesis deals with a new approach to construction of confidence regions for multivariate random variables and multivariate random samples. This can also be viewed as one of the possible generalizations of the notion of quantile into a multidimensional case. The approach is based on the following: in the first step, a centred random vector is transformed into polar (hyperspherical) coordinates. Afterwards, so-called directional quantiles are determined. These are classical unidimensional quantiles for distribution of the radius conditional on the angle of the polar coordinates. Sample analogy of the directional quantiles is estimated using trigonometrical series with coefficients obtained by quantile regression. The first chapter deals with the choice of the origin for the centralization of the data. We examine both theoretical and sample cases. We offer several variants with focus on the deepest point. The second chapter concerns quantile regression with focus on the aspects, which have an impact on the properties of sample periodical regression quantiles. The third and most exhaustive chapter is devoted to periodical regression quantiles construction and properties. Both theoretical and sample variants and their relationship are described. Several examples are offered in the end of the chapter.
|
|
| |
|
| |
|
| |
|
|
Unfair Ballots
Valášková, Zuzana ; Lachout, Petr (referee) ; Hlubinka, Daniel (advisor)
OBSAH Nazov prace: Nespravodlivc losovanie Antor: Zuzana Valaskova Katedra: Katedra pravdepodobnosti a matematicke statistiky Veduci bakalarskej prace: R.NDr.Daniel Hlnbinka, Ph.D. e-mail veduceho pracc: hluhu]k;i:'"Jkarlm.mif.cuni.cz Abstrakt: V predlozencj praci studujem otazku ncspravodliveho losovania, ktorn som sku- mala prostrednictvom silneho nastroja, testovania hypotez. Na zaciatku som sa zaoberala najdenim vhodnej nahodnej veliciny. ktora by iiam umoznila jedno- duchym sposobom rozrieait! povodnu otazkn spravodlivosti losovania. Zaoberala som sa otazkou randomizovanych a nerandomizovanych testov. V praci su okrein ineho nvedenc iri vzorove priklady s podrobnyrn riescnim. Vo vsetkych jc nvedena aplikacia randoinizovaneho i nerondomizovaneho testu. Prvy ilustruje prave take losovaiiie, u ktoreho neniame dostatocne dovody na pochybovanie o spravodlivosti celeho losovania, v drnhoni priklade je uvedeny typicky pri[)ad ncspravodliveho losovania a v l.rel'oTn je uvedeny pripa.fi. kedy t,(^sty nedavaju rovnake vyskxlky. Prilohu tvori zdrojovy text nnmerickych vypoctov v Mathernatice. Kmmve slova: losovanie, sf>rav(>dlivost'.tnst Title: Unfair Ballots Author: Zuzana Valaskova Department: Department of Probability and Mathematical Statistics Supervisor: RNDr.Daniel Illubinka, Ph.D. Supervisor's e-mail...
|
|
| |
|
| |
|
| |
|
|
The identification function for the convergence in probability with an application in the estimation theory
Kříž, Pavel ; Hlubinka, Daniel (referee) ; Štěpán, Josef (advisor)
In the present work we introduce the concept of probability limit identification function (PLIF) as it is done in [6]. This function identifies almost surely the value of the probability limit of a sequence of random variables on the basis of one realization of the sequence. According to the same article we show the construction of PLIF for real valued random variables from the special PLIF for 0-1 valued random variables. Following the method described in [8] we prove the existence of the universal PLIF for real valued random variables under the continuum hypothesis. Next we show that there are no borel measurable special PLIFs for 0-1 valued random variables (as well as PLIFs for real valued random variables). We use the proof that is published in [2]. Then we extend the construction of PLIF from R to any separable metrizable topological space. This PLIF may be used e.g. for creating functional representations of stochastic integrals and weak solutions of stochastic differential equations.
|
guest ::
RSS feed.