|
|
Essays on Data-driven, Non-parametric Modelling of Time-series
Hanus, Luboš ; Vácha, Lukáš (advisor) ; Witzany, Jiří (referee) ; Ellington, Michael (referee) ; Trimborn, Simon (referee)
This thesis consists of four contributions to the literature on data-driven and non-parametric modelling of time series. In the first paper, we study the synchronisation of business cycles and propose a multivariate co-movement measure based on time-frequency cohesion. We suggest that economic inte- gration may lead to increased co-movement of business cycles, which may reflect the benefits of convergence and coordination of economic policies. The second paper presents a new methodology for identifying persistence in macroeconomic variables. Using time-varying frequency response func- tions, we identify heterogeneous persistence effects in US macroeconomic variables. The third and fourth papers propose data-driven techniques for probabilistic forecasting of time series using deep learning. We introduce a multi-output neural network that selects the most appropriate distribution for the data. The distributional neural network is valuable for modelling data with non-linear, non-Gaussian and asymmetric structures. The third paper demonstrates the usefulness of the method by estimating information-rich macroeconomic fan charts and distributional forecasts of asset returns. In the last paper, we present the distributional neural network to obtain the proba- bility distribution of electricity price...
|
|
|
Digital Image Steganography
Mazánek, Antonín ; Šimek, Václav (referee) ; Strnadel, Josef (advisor)
This bachelor's thesis introduces the steganography scientific discipline. It presents its ancient and relative recent history, mentions its current use and outlines the possible directions for steganography future. It introduces steganography into context of hiding information superior domain and highlights differences between other methods in this domain. It describes steganography subdivisions and closer deals with digital image steganography. It presents some known methods and types of digital image data. The thesis also contains more detailed description of image data in PNG format. It describes the implementation of LSB and PVD steganography methods with images in PNG format as cover object and evaluates their key properties, namely: imperceptibility, capacity and robustness. At the end of the thesis, the author's reflection on the results and possibilities of further modifications of the mentioned methods is given.
|
|
|
Arc Fault Detection
Fendrychová, Michaela ; Chumak,, Oleksiy (referee) ; Orságová, Jaroslava (advisor)
Diplomová práce je zaměřena na problematiku analýzy signálů za účelem detekce poruchového oblouku, přičemž analýza signálů je prováděna v časové, frekvenční a smíšené časově-frekvenční oblasti. Práce stručně shrnuje existující normy pro zařízení pro detekci poruchového oblouku. Práce dokumentuje testy a měření, které byly realizovány v souladu s normami IEC 62606:2013 a UL 1699B. Z důvodu nedostatečnosti stávajících norem je v práci popsána nová metoda iniciace poruchového oblouku. naměřená data byla analyzována s využitím rychlé Fourierovy transformace, krátkodobé Fourierovy transformace a vlnkové transformace. Na základě provedeného literárního průzkumu a s využitím výsledků provedených analýz signálu je v práci proveden návrh nové detekční metody pro účely detekce poruchového oblouku v systémech napájených střídavým i stejnosměrným napětím.
|
|
|
Time domain edge finite elements
Cigánek, Jan ; Motl, Milan (referee) ; Raida, Zbyněk (advisor)
Diplomová práce se zabývá metodou hybridních (hranových a uzlových) konečných prvků ve frekvenční i časové oblasti. Tato metoda je použita pro analýzu vlnovodu parallel-plate, v kterém jsou umístěny dvě dielektrické vrstvy. Jako ukončení vlnovodu je implementována dokonale přizpůsobená vrstva označována PML. Projekt řeší možný výběr PML vrstvy v časové oblasti. Metoda je programována v programu MATLAB a výsledky jsou porovnány s programem COMSOL Multiphysics.
|
|
| |
|
|
Removal of a Known Signal from a Recording
Urbánek, Radomír ; Žmolíková, Kateřina (referee) ; Černocký, Jan (advisor)
The goal of this diploma thesis is to design and implement method for removing known signal from recorded sound. In the introductory part of the thesis are described the properties of sound and its propagation through the space, digital signal processing on the computer and the computing of the system impulse response. The sequential development of the methods leading to the removal of the known signal from the recording is described further. The following part contains a description and implementation of this method. It also describes how and at what data the method will be tested. Finally, testing is evaluated, improvements and further possible work is proposed.
|
|
|
Analysis of Heart Rate Variability
Škrtel, Karol ; Kolářová, Jana (referee) ; Provazník, Ivo (advisor)
The project describes the methods useful for observe changes of heart rate in ECG signal. Heart rate variability become (HRV) the conventionally accepted term to describe variations of NN intervals between consecutive heart beats and generally it is function of instantaneous heart rate or NN interval on time. HRV may be evaluated by time domain or frequency domain measures. In Matlab was developed algorithm, realized like function, which counts HRV parameters from ECG signal series. Analysis in time domain adverts to high correlation between statistic and geometric parameters and similarly with signal HRV. Results of frequency domain analysis shows similarity of power spectral density, which was calculated by two different ways (from interpolated and no interpolated signal HRV). Functionality of developed algorithm was verified on each signal. Project results have signification in progress of analysis ECG signal methods with a view to observe pathological changes in heart rate.
|
|
|
Detection of gunshots from small arms
Nesvadba, Ondřej ; Malucha, Jan (referee) ; Sigmund, Milan (advisor)
This thesis deals with acoustical gunshot detection from small arms, typically of the calibre up to 10 mm, primarily in urban areas. Thesis includes initial research of methods, which are usually used for gunshot detection. The key part of this work is examination of typical values of gunshot signal. Based on the values obtained, decisive parameters in time and frequency domains are determined and detection algorithms are proposed. Function of these algorithms is tested and evaluated. The possible influence of various features of recordings on detection success rate was discussed.
|
|
|
Music Source Separation
Holík, Viliam ; Veselý, Karel (referee) ; Mošner, Ladislav (advisor)
Neural networks are used for the problem of music source separation from recordings. One such network is Conv-TasNet. The aim of the work is to experiment with the already existing implementation of this network for the purpose of potential improvement. The models were trained on the MUSDB18 dataset. It was successively experimented with the change of the network structure, transforming signals from the time domain to the frequency domain for the purpose of calculating the loss function, replacing different loss functions with the original one, finding the optimal learning rate for each loss function and gradually decreasing the learning rate during the learning process. The best experiments according to the SDR metric were training with loss functions L1 and logarithmic L2 in the time domain with a higher initial learning rate with its gradual decrease during the learning process. In a relative comparison of the best models to the baseline, it is more than 2.5% improvement.
|
|
|
Sound Effect "Glitch Machine"
Cimmerman, Matej ; Mošner, Ladislav (referee) ; Černocký, Jan (advisor)
This thesis addresses software music effects and their implementation. It shows design and implementation of the Glitch Machine, which belongs to the music multieffect programs and demonstrates the simultaneous use of various music effects to process audio signals. The result is a program for Windows OS, which utilizes cooperation of various effects, namely a filter, reverb, distortion, extractor, reverz, stutter, shifter, pitch resample and gain. It allows extensive and detailed editing of an audio track and a higher extent of manipulation with sounds, than with single music effects. Other components of the program include visualisations of the waveform and frequency spectrum of audio signals and an option for loading and saving effect presets. The program was tested in tests of the performance and memory usage and also with user testing of quality and graphical interface.
|
guest ::
