|
|
Cluster analysis for functional data
Zemanová, Barbora ; Komárek, Arnošt (advisor) ; Hušková, Marie (referee)
In this work we deal with cluster analysis for functional data. Functional data contain a set of subjects that are characterized by repeated measurements of a variable. Based on these measurements we want to split the subjects into groups (clusters). The subjects in a single cluster should be similar and differ from subjects in the other clusters. The first approach we use is the reduction of data dimension followed by the clustering method K-means. The second approach is to use a finite mixture of normal linear mixed models. We estimate parameters of the model by maximum likelihood using the EM algorithm. Throughout the work we apply all described procedures to real meteorological data.
|
|
|
Classification based on longitudinal observations
Bandas, Lukáš ; Komárek, Arnošt (advisor) ; Kulich, Michal (referee)
The concern of this thesis is to discuss classification of different objects based on longitudinal observations. In the first instance the reader is introduced to a linear mixed-effects model which is useful for longitudinal data modeling. Description of discriminant analysis methods follows. These methods ares usually used for classification based on longitudinal observations. Individual methods are introduced in the theoretic aspect. Random effects approach is generalized to continuous time. Subsequently the methods and features of the linear mixed-effects model are applied to real data. Finally features of the methods are studied with help of simulations.
|
|
|
Regression models with alternatively distributed response
Kučera, Tomáš ; Komárek, Arnošt (advisor) ; Zvára, Karel (referee)
This thesis deals with regression models in the case of binary response variable. Linear and logistic regression models are defined for different types of predictors. Then the thesis uses the theory of maximum likelihood and applies it to the special case of logistic regression model. Both exact inference of model parameters and hypothesis testing with related interval inference are discussed. Suitable methods for numerical solving of selected methods are suggested. In the final part, the discussed methods are applied to real credit scoring data from the field of banking, using the statistical software R.
|
|
|
Simpson's paradox
Balhar, Jan ; Komárek, Arnošt (advisor) ; Kulich, Michal (referee)
Title: Simpson's paradox Author: Jan Balhar Department: Department of Probability and Mathematical Statistics Supervisor: RNDr. Arnošt Komárek, Ph.D. Supervisor's e-mail address: arnost.komarek@mff.cuni.cz Abstract: In this work we deal with Simpson's paradox and its more general version, called association reversal. We present definitions of these terms and necessary and sufficient conditions for their occurrence. Due to this, we get to issue of measuring relationship between two characters in 2x2 contigency table, we specifically mention advantages of odds ratio. We also try to answer, what relationship between two characters is, in case of Simpson's paradox, the right one. When looking for answer, we find, that ordinary statistical methods are not sufficient. It is necessary to identify causal relationships between characters. Therefore we get to issue of causality definition. Finally, we present some examples of Simpson's paradox in practice. Keywords: Simpson's paradox, association reversal, confounding, causality.
|
|
| |
|
|
Some problems of exponential smoothing
Čurda, David ; Hanzák, Tomáš (advisor) ; Komárek, Arnošt (referee)
In this work the several exponential smoothing type methods are briefly described, which are often used to smoothing and forecasting in the time series. Selected problems, that occur in described methods, are presented and in some cases there are the suggestions to their solution, which should tend to more suitable smoothing or to the better forecasts. It's shown how the methods are applied on different data and how the forecasts differ from each other. In conclusion the quality of modifications is evaluated.
|
|
|
Computational Methods for Maximum Likelihood Estimation in Generalized Linear Mixed Models
Otava, Martin ; Komárek, Arnošt (advisor) ; Kulich, Michal (referee)
of the diploma thesis Title: Computational Methods for Maximum Likelihood Estimation in Generalized Linear Mixed Models Author: Bc. Martin Otava Department: Department of Probability and Mathematical Statistics Supervisor: RNDr. Arnošt Komárek, Ph.D., Department of Probability and Mathematical Statistics Abstract: Using maximum likelihood method for generalized linear mixed models, the analytically unsolvable problem of maximization can occur. As solution, iterative and ap- proximate methods are used. The latter ones are core of the thesis. Detailed and general introducing of the widely used methods is emphasized with algorithms useful in practical cases. Also the case of non-gaussian random effects is discussed. The approximate methods are demonstrated using the real data sets. Conclusions about bias and consistency are supported by the simulation study. Keywords: generalized linear mixed model, penalized quasi-likelihood, adaptive Gauss- Hermite quadrature 1
|
|
|
Errors in Variables
Mordinová, Katarína ; Hlávka, Zdeněk (advisor) ; Komárek, Arnošt (referee)
1 Title: Errors in variables Author: Katarína Mordinová Department: Department of Probability and Mathematical Statistics Supervisor: Mgr. Zdeněk Hlávka, Ph.D. Supervisor's e-mail address: Zdenek.Hlavka@mff.cuni.cz Abstract: The topic of the diploma thesis is Errors in variables. In the opening chapter, we define basic terms used in the thesis and we introduce the regression analysis and basic relations related to this term. In the second chapter, we attend to linear regression model and its characteristics. In the third chapter, we attend to the errors in variables models. In the last chapter of this thesis we present a possible application in medicine. Keywords: regresion analysis, errors in variables, linear regression model
|
|
| |
|
| |
guest ::
RSS feed.