The Contracts Register is a public database containing contracts concluded by public institutions. Due to the number of documents in the database, data analysis is proble- matic. The objective of this thesis is to find a machine learning approach for sorting the contracts into categories by their area of interest (real estate services, construction, etc.) and implement the approach for usage on the web portal Hlídač státu. A large number of categories and a lack of a tagged dataset of contracts complicate the solution. 1

The statistical distribution of daily changes in temperature is not symmetrical, and also the day-to-day temperature changes distribution varies for winter and summer. The aim of submitted thesis is to connect the asymmetry of distribution with physical causes. Specifically, the aim is to check the following hypothesis: the more frequent occurrence of coolings up to -3 řC in winter a heatings up to 3 řC in summer is associated with radiation balance during the anticyclonic flow over Central Europe. The relationship of day-to-day temperature changes with passages of atmospheric fronts is also described. Temperature changes in winter (December - February) are calculated from minimum temperatures while in summer (June - August) maximum temperatures are used. The study period covers years between 1961 and 1998. The Hess-Brezowsky catalogue of weather types, the HMÚ classification and selected classifications of COST733 database are used for description of circulation conditions in Central Europe. The connection of anticyclonic flow with frequent coolings in winter was not proved. On the other hand, in summer, the association between frequent warmings and anticyclonic flow was confirmed. In addition, the advection of cold air in winter and warm air in summer also contributes to asymmetry of the...

Tick-borne encephalitis risk assessment based on classification of vegetation from remote sensing data Abstract The main aim of this thesis has been to find out how to classify various categories of forest vegetation with a different risk of exposure to the tick-borne encephalitis based on the Landsat imagery. The legend used here is derived from the one used in the projects by Daniel, Kolář, Zeman (1995) and Daniel, Kolář, Beneš (1999) but has been reduced to only five classses with no overlaps in their definitions (I. coniferous stands, II. mixed stands, III. young deciduous stands and stand ecotones with a highly heterogeneous structure, IV. deciduous stands with a homogeneous structure, V. deciduous stands with a heterogeneous structure). The supervised classification with the Maximum Likelihood Classifier has been used on the Landsat imagery from various seasons. Difficulties concerned with the presence of clouds and varying Sun elevation across the imagery had to be adressed in the course of the work. The training sites and the control points have been defined by the field research and interpretation of the relevant orthophotomaps and Landsat imagery in 5-4-3 RGB composite. The mask of the forest has been created on the ZABAGED data basis. The time horizon of 2006 - 2010 has been the primary focus....

The thesis deals with use of hyperspectral data from APEX and AISA sensors for detection and classification of anthropogenic materials in the areas of Čáslav, Rokytnice nad Jizerou and Harrachov. The main goal is to propose methodology for the detection and classification of roof materials and road surface materials based on established spectral libraries. Another goal is to evaluate applicability of spectral libraries for classification, to compare possibilities of hyperspectral data with larger and smaller spectral range and to create maps of anthropogenic materials above. The methodological approach including masks of anthropogenic materials for roads surface materials and roof materials creation, settings of four classifications algorithms (Linear Spectral Unmixing, Multiple endmember spectral mixture analysis, Spectral Angle Mapper, Spectral Information Divergence) parameters and assessment of classification results, is in the methodology part. The results are visualized and evaluated using overall accuracy and percentage of classified pixels. Finally the results are compared with existing studies and possible improvements for further research are proposed. Powered by TCPDF (www.tcpdf.org)

Financial instruments stand for a critical role in current economy. Total value of financial assets was $167 trillion due to McKinsey study in 2007. More alarming is a fact that total value of all derivative at OTC market was more than $600 trillion what is more than world's wealth. It is obvious that such a market needs rigorous regulation. International Accounting Standard Board (IASB) releasing new standard which regulate only financial instruments aims to react on that fact. In first part I would like to show current situation on financial market and brief regulation due to current rules. From the text there should be obvious nowadays turbulent environment and crisis deforming financial market. It was necessity for me to demonstrate the process of developing and adopting of the new standard. Next part of my diploma thesis refers to adopted rules for financial assets. This should be the fundamental part for users of financial statements as well as for my diploma thesis. The main changes took part in classification of financial assets. There are two new concepts for classification of financial asset which should be described in detailed. For better illustration I also placed practical part for classification according to new models. For readers it is convenient to know the differences between exposure draft and released standard. That should demonstrate the complex view of IASB. There are few changes in classification of financial liabilities. The main change is in specific case of own credit which I posted in thesis. This specific problem postponed the whole endorsement of the standard. The last chapter is dedicated to convergence between IFRS and US GAAP. Both parties work on convergence on reporting of financial instruments.

3D Object Classification Using Neural Networks Bc. Miroslav Krabec Classification of 3D objects is of great interest in the field of artificial intelligence. There are numerous approaches using artificial neural networks to address this problem. They differ mainly in the representation of the 3D model used as input and the network architecture. The goal of this thesis is to explore and test these approaches on publicly available datasets and subject them to independent comparison, which has not so far appeared in the literature. We provide a unified framework allowing to convert the data from common 3D formats. We train and test ten different network on the ModelNet40 and ShapeNetCore datasets. All the networks performed reasonably well in our tests, but we were generally unable to achieve the accuracies reported in the original papers. We suspect this could be due to extensive, albeit unreported, hyperparameter tuning by the authors of the original papers, suggesting this issue would benefit from further research. 1

This work describes contemporary methods for sound data classification and their application on sound recordings. This work deals with the selection of sound signal characteristics appropriate for concrete classification criteria followed by the exploration of proper types of artificial neural networks for this task. Different topologies and learnings algorithms of chosen neural networks are proposed and their performance in this classification is compared. Results compared are discussed and the best solution for sound data classificaton based on the content of recordings is chosen. This work also includes the chosen algorithms implementation into the software solution of my own.

Increasing interest for classification of 3D geometrical data has led to discov- ery of PointNet, which is a neural network architecture capable of processing un- ordered point sets. This thesis explores several methods of utilizing conventional point features within PointNet and their impact on classification. Classification performance of the presented methods was experimentally evaluated and com- pared with a baseline PointNet model on four different datasets. The results of the experiments suggest that some of the considered features can improve clas- sification effectiveness of PointNet on difficult datasets with objects that are not aligned into canonical orientation. In particular, the well known spin image rep- resentations can be employed successfully and reliably within PointNet. Further- more, a feature-based alternative to spatial transformer, which is a sub-network of PointNet responsible for aligning misaligned objects into canonical orientation, have been introduced. Additional experiments demonstrate that the alternative might be competitive with spatial transformer on challenging datasets. 1

The following thesis explores the impact of the dataset distributional prop- erties on classification performance. We use Gaussian copulas to generate 1000 artificial dataset and train classifiers on them. We train Generalized linear models, Distributed Random forest, Extremely randomized trees and Gradient boosting machines via H2O.ai machine learning platform accessed by R. Classi- fication performance on these datasets is evaluated and empirical observations on influence are presented. Secondly, we use real Australian credit dataset and predict which classifier is possibly going to work best. The predicted perfor- mance for any individual method is based on penalizing the differences between the Australian dataset and artificial datasets where the method performed com- paratively better, but it failed to predict correctly. 1

No description available.