|
Solving image analysis problems by minimization of total variation
Janáček, Jiří
A solution of various problems in image analysis using concurrent minimization of total variation and Lp loss function is presented. The minimization is achieved by a steepest descend method using graph cut minimization in each step. Regularization of noisy image and registration of microscopic images of physical sections are demonstrated
|
|
Testování homogenity a dobré shody v analýze přežití
Timková, Jana
The present paper deals with the goodness of fit and the twosample problem related to the event-history type data. The proposed methods are derived from bayesian nonparametric approach and take advantage of MCMC estimation of the hazard rate. The technique is based on Bayes construction of martingale residuals.
|
|
Zobecněný Zero range proces jako model toku dopravy
Fajfrová, Lucie
In the paper, a model of conservative particle system, which generalises a well known zero range process, is studied. The generalisation consists in allowing jumps of more than one particle in one moment. We describe what this generalisation means in the context of modeling a traffic flow.
|
|
O významu entropie
Janžura, Martin
The aim of the paper consists in demonstrating the relevance of the fundamental information-theoretic concepts, namely the entropy and the I-divergence, for both the statistical inference and the limit theorems of probability theory.
|
|
O odhadu vzájemné informace
Marek, Tomáš ; Tichavský, Petr
The mutual information is useful measure of a random vector component dependence. It is important in many technical applications. The estimation methods are often based on the well known relation between the mutual information and the appropriate entropies. In 1999 Darbellay and Vajda proposed a direct estimation methods. In this paper we compare some available estimation methods using different 2-D random distributions.
|
| |
|
Determination of the Optimal Number of Clusters in Statistical Software Systems
Řezanková, H. ; Húsek, Dušan
The paper deals with approaches to determination of the optimal number of groups of objects which are implemented in clustering algorithms in commercial statistical software packages. Some of these approaches are applied to the example of finding groups of deputies of the Russian parliament on the base of roll-call votes in 2004. The capabilities of software packages S-PLUS, SAS, SPSS, and SYSTAT are described.
|
| |
|
Interval Estimate of Binomial Parameter p: What is (Relatively) New?
Klaschka, Jan
The present work follows up the ROBUST 2006 paper where various types of confidence intervals for binomial parameter p have been exposed. The coverage probability cannot equal the nominal confidence level 1-alpha in the whole domain [0, 1]. This leads to dilemmas (is the coverage of at least 1-alpha a must, or is it better to approximate 1-alpha from both sides?), and to multiplicity of proposals of confidence interval types. The present work extends the scope of the previous paper by such generalizations of "ordinary" confidence intervals that enable a constant coverage, namely by the randomized confidence intervals (introduced several decades ago), and by the relatively new idea of the fuzzy confidence intervals.
|