|
|
Binning numerical variables in credit risk models
Mattanelli, Matyáš ; Baruník, Jozef (advisor) ; Teplý, Petr (referee)
This thesis investigates the effect of binning numerical variables on the per- formance of credit risk models. The differences are evaluated utilizing five publicly available data sets, six evaluation metrics, and a rigorous statistical test. The results suggest that the binning transformation has a positive and significant effect on the performance of logistic regression, feedforward artifi- cial neural network, and the Naïve Bayes classifier. The most affected aspect of model performance appears to be its ability to differentiate between eligible and ineligible customers. The obtained evidence is particularly pronounced for moderately-sized data sets. In addition, the findings are robust to the inclusion of missing values, the elimination of outliers, and the exclusion of categorical features. No significant positive effect of the binning transformation was found for the decision tree algorithm and the Random Forest model.
|
|
|
Simulation of fluid flow around obstacles by Lattice Boltzmann Method
Prinz, František ; Pokorný, Jan (referee) ; Zatočilová, Jitka (advisor)
The task of this diploma thesis is the Lattice Boltzmann method (LBM). LBM is a mesoscopic method describing the particle motion in a fluid by the Boltzmann equation, where the distribution function is involved. The Chapman-Enskog expansion shows the connection with the macroscopic Navier-Stokes equations of conservation laws. In this process the Hermite polynoms are used. The Lattice Boltzmann equation is derived by the discretisation of velocity, space and time which is concluding to the numerical algorithm. This algorithm is applied at two problems of fluid flow: the two-dimensional square cavity and a flow arround obstacles. In both cases were the results of velocities compared to results calculated by finite volume method (FVM). The relative errors are in order of multiple 1 %.
|
|
|
Commented translation of a text on science and technology
Heger, Matěj ; Krhutová, Milena (referee) ; Zmrzlá, Petra (advisor)
Bakalářská práce se zaměřuje především na překlad odborného textu s důrazem na použitou terminologii. Vybraný text pojednává o problematice mikrovln a jejich zpracování v počítačových systémech. Práce sestává z dvou stěžejních částí, přičemž první se soustřeďuje na výše zmíněný překlad odborného textu a druhá se věnuje jeho podrobné jazykové analýze. Analytický rámec zahrnuje všechny důležité jazykovědné disciplíny a předkládá výňatky přímo z textu. Tyto výňatky jsou následně porovnávány a postupně je ověřována správnost použitých jazykových prostředků pomocí citací z kvalitních zdrojů a slovníků.
|
|
|
Inequalities for discrete and continuous supremum operators
Oľhava, Rastislav ; Pick, Luboš (advisor) ; Cianchi, Andrea (referee) ; Nekvinda, Aleš (referee)
Inequalities for discrete and continuous supremum operators Rastislav O , lhava In this thesis we study continuous and discrete supremum operators. In the first part we investigate general properties of Hardy-type operators involving suprema. The boundedness of supremum operators is used for characterization of interpo- lation spaces between two Marcinkiewicz spaces. In the second part we provide equivalent conditions for boundedness of supremum operators in the situation when the domain space in one of the classical Lorentz spaces Λp w1 or Γp w1 and the target space Λq w2 or Γq w2 . In the case p ≤ q we use inserting technique obtaining continuous conditions. In the setting of coefficients p > q we provide only partial results obtaining discrete conditions using discretization method. In the third part we deal with a three-weight inequality for an iterated discrete Hardy-type operator. We find its characterization which enables us to reduce the problematic case when the domain space is a weighted ℓp with p ∈ (0, 1) into the one with p = 1. This leads to a continuous analogue of investigated discrete inequality. The work consists of author's published and unpublished results along with material appearing in the literature.
|
|
|
Inequalities for discrete and continuous supremum operators
Oľhava, Rastislav ; Pick, Luboš (advisor) ; Nekvinda, Aleš (referee)
Inequalities for discrete and continuous supremum operators Rastislav O , lhava In this thesis we study continuous and discrete supremum operators. In the first part we investigate general properties of Hardy-type operators involving suprema. The boundedness of supremum operators is used for characterization of interpo- lation spaces between two Marcinkiewicz spaces. In the second part we provide equivalent conditions for boundedness of supremum operators in the situation when the domain space in one of the classical Lorentz spaces Λp w1 or Γp w1 and the target space Λq w2 or Γq w2 . In the case p ≤ q we use inserting technique obtaining continuous conditions. In the setting of coefficients p > q we provide only partial results obtaining discrete conditions using discretization method. In the third part we deal with a three-weight inequality for an iterated discrete Hardy-type operator. We find its characterization which enables us to reduce the problematic case when the domain space is a weighted ℓp with p ∈ (0, 1) into the one with p = 1. This leads to a continuous analogue of investigated discrete inequality. The work consists of author's published and unpublished results along with material appearing in the literature.
|
|
|
The effect of sampling period error on frequency analysis
Pietrowicz, Daniel ; Marada, Tomáš (referee) ; Zuth, Daniel (advisor)
This thesis will be focused on signal measurement in connection with digitization and analysis inaccuracy of sampling peroid effect. Brief narration of frequency analysis and description of Matlab and his capabilities with this issue. Then creation procedure for simulation model of analog signal and his fault digitization via sampling period. Systematic comparation of this effect on inaccuracy of sampling period on frequency analysis.
|
|
|
Probabilistic Spacetimes
Káninský, Jakub ; Svítek, Otakar (advisor) ; Žofka, Martin (referee)
Probabilistic Spacetime is a simple generalization of the classical model of spa- cetime in General Relativity, such that it allows to consider multiple metric field realizations endowed with probabilities. The motivation for such a generalization is a possible application in the context of some quantum gravity approaches, na- mely those using the path integral. It is argued that this model might be used to restrict the precision of the geometry on small scales without postulating discrete structure; or it may be used as an effective description of a probabilistic geometry resulting from a full-fledged quantum gravity computation.
|
|
|
Recursive calculation of compound distributions
Sekera, Michal ; Mazurová, Lucie (advisor) ; Pešta, Michal (referee)
The aim of this work is the calculation of compound distributions by using the algorithm known as the Panjer recursion. This algorithm is limited to dis- crete distributions and the (a, b, 0) and (a, b, 1) classes of distributions. The thesis shows which distributions are members of these classes. The thesis then descri- bes the discretization of continuous distributions by using the rounding method and the method of local moments matching; everything is explained on exam- ples. These methods are then applied to the calculation of the premium for mo- del of excess of loss reinsurance with reinstatements (XL-reinsurance with re- instatements), and the calculating the solvency capital requirement. Numerical illustrations are included. 1
|
|
|
Web Page Classification
Kolář, Roman ; Burget, Radek (referee) ; Bartík, Vladimír (advisor)
This paper presents problem of automatic webpages classification using association rules based classifier. Classification problem is presented, as a one of datamining technique, in context of mining knowledges from text data. There are many text document classification methods presented with highlighting benefits of classification methods using association rules. The main goal of work is adjusting selected classification method for relation data and design draft of webpages classifier, which classifies pages with the aid of visual properties - independent section layout on the web page, not (only) by textual data. There is also ARC-BC classification method presented as a selected method and as one of intriguing classificators, that derives accuracy and understandableness benefits of all other methods.
|
|
|
Preprocessing Methods in Knowledge Discovery
Pokorný, Lubomír ; Lukáš, Roman (referee) ; Bartík, Vladimír (advisor)
This thesis discusses preprocessing methods in knowledge discovery. There are four methods of discretization, filling in the missing values and normalization is investigated. There is closer view on building the application which discretizate numeric atribute in database. All methods were successfully tested and the results were added.
|
guest ::
