National Repository of Grey Literature 543 records found  1 - 10nextend  jump to record: Search took 0.00 seconds. 
Deep Neural Networks for Time Series Forecasting
Kayabasi, Yigit Mertol ; Pilát, Martin (advisor) ; Neruda, Roman (referee)
Time series forecasting is a task of both academic and pragmatic interest. Although it has been long dominated by qualitative methods and simple quan- titative methods, machine learning and deep learning algorithms in modelling temporal data has become more common, but the progress is still far from the progress in typical machine learning tasks like computer vision or natural lan- guage processing. Recurrent neural networks are the most natural choice for modelling sequential data, but training them is tricky especially to learn from long sequences. Recently a divergence from back propagation Reservoir Comput- ing paradigm has started to draw attention with the performance of the models arising from it in this kind of tasks. They proved to be a good option partic- ularly for modelling rather more chaotic systems. In this thesis we will explore and compare these two families of neural networks regarding their performance and implementation. 1
Three essays on empirical Bayesian econometrics
Adam, Tomáš ; Komárek, Luboš (advisor) ; Feldkircher, Martin (referee) ; Herrala, Risto (referee) ; Melecký, Martin (referee)
The dissertation consists of three papers which apply Bayesian econometric techniques to monitoring macroeconomic and macro-financial developments in the economy. Its aim is to illustrate how Bayesian methods can be employed in standard areas of economic research (estimating systemic risk in the banking sectors, nowcasting GDP growth) and also in a more original area (monitoring developments in sovereign bond markets). In the first essay, we address a task which analytical departments in central banks or commercial banks face very often - nowcasting foreign demand of a small open economy. On the example of the Czech economy, we propose an approach to nowcast foreign GDP growth rates for the Czech economy. For presentation purposes, we focus on three major trading partners: Germany, Slovakia and France. We opt for a simple method which is very general and which has proved successful in the literature: the method based on bridge equation models. A battery of models is evaluated based on a pseudo-real- time forecasting exercise. The results for Germany and France suggest that the models are more successful at backcasting, nowcasting and forecasting than the naive random walk benchmark model. At the same time, the various models considered are more or less successful depending on the forecast horizon....
Modelling Duration of Financial Transaction Data
Nácovský, Patrik ; Hendrych, Radek (advisor) ; Branda, Martin (referee)
This bachelor thesis deals with ACD (autoregressive conditional duration) model, which is used to estimate durations of time series of financial transaction data. First, duration and time series are defined formally as well as with the intuitive way. Next, model ACD itself is defined and its basic types, which are determined with distribution of its residuals. Then way to use this model for predictions is introduced. In the second part, steps for model identification, construction and revision are described. In the last part models EACD, WACD and GACD are constructed for real data. There are three data sets of thick data, which are Apple stocks, EUR/USD and gold. Data sets contain from 300 thousands to 600 thousands elements (one trading week).
Normalization of Time Series Data of Landsat
Svoboda, Jan ; Štych, Přemysl (advisor) ; Kolář, Jan (referee)
Spectral reflectance of the Earth surface, obtained from the satellite images, should be independent from the external influences and should reflect the surface properties, specifically the proportion of the radiance reflected from the object. It was proved in this paper that the time series of the 63 images from the Landsat 5 satellite were visibly influenced by the external factors even in the case of the images already atmospherically corrected. These external factors were age of the image and WRS-2 position from which the image was obtained. Age of the image was documented with the steady decrease of the spectral reflectance values of the invariant features, especially in the visible part of the electromagnetic spectrum, caused by the sensor degradation. The influence of the WRS-2 position was documented especially in the infrared bands. The western parts of the images are lighter (have higher values of the surface reflectance) than the eastern parts. That may cause the difference between values when monitoring one spot in two overlapping WRS-2 positions. The method originally used for the relative radiometric normalization IR-MAD was here applied to normalize the surface reflectance data, and resulted in the fact that these influences did not show up any more. In order to extend the time...
Predictive Modelling with Python
Duda, Jan ; Burgetová, Ivana (referee) ; Zendulka, Jaroslav (advisor)
The main goal of this bachelor thesis is get to know with the data mining and its domain, also with the Knowledge discovery in databases process. It shows the most importnant approaches, which are implemented in Python language afterwards. The case study contains the prediction of index S&P 500 describing stock market developments on the US stock exchange. Both classification and regression models are used for the forecasting. Model evaluation is reached by the Monte Carlo experimental method.
Analysis of Network Security Alerts
Dobeš, Erik ; Žádník, Martin (referee) ; Wrona, Jan (advisor)
The goal of this work is to find groups of IP addresses in network security reports, which were detected in the same, or very similar, time interval. The work introduces an algorithm, which transforms data from security reports into time series. Between all the time series, similar pairs are searched. Subsequently, in the found pairs, we are looking for similar threesomes, in which we try to find similar foursomes, etc. The created solution successfully found 208 similar groups in the set of analyzed data, the largest of which contains 11 similar IP addresses. Based on the data found it is possible to detect machines that are part of the so-called botnet in network security reports.
Anomaly Detection in Generated Incident Ticket Volumes
Šurina, Timotej ; Rychlý, Marek (referee) ; Trchalík, Roman (advisor)
Táto bakalárska práca sa zaoberá problematikou detekcie anomálií v časových radoch. Predstavuje metódy STL decomposition, ARIMA, Exponential Smoothing a LSTM Networks. Cieľom je pomocou týchto metód vytvoriť algoritmus, ktorý dokáže analyzovať trend v množstve generovaných záznamov o incidentoch a detekovať anomálie z trendu. Riešenie bolo vytvorené na základe dátovej sady poskytnutej firmou AT&T Global Network Services Czech Republic s.r.o. a implementované v programovacom jazyku Python.
Statistical Modeling of the Risk Indicators in a Company
Rufer, Jiří ; Žák, Libor (referee) ; Karpíšek, Zdeněk (advisor)
This thesis aims to analyze accounting and financial indicators using time series methods and interval regression analysis for Rudolf Jelínek, a.s. In this thesis are analyzed development trends of individual indicators. Based on the obtained data, the company deals with the risks of the company based on analyzes and their solutions.
Analysis of Economic Indicators using Statistical Methods
Matoušek, Ondřej ; Linhart, Michal (referee) ; Michalíková, Eva (advisor)
This bachelor thesis focuses on the calculation and subsequent evaluation of economic indicators of the chosen company using financial analysis and statistical methods for predicting the future development of selected indicators. The thesis presents suggestions for improving the financial situation of the analyzed company. The output of the bachelor thesis is also the VBA computer program that allows automatically calculate the selected financial indicators based on the data from the financial statements.

National Repository of Grey Literature : 543 records found   1 - 10nextend  jump to record:
Interested in being notified about new results for this query?
Subscribe to the RSS feed.