|
|
Stock Markets Analysis Using New Genetic Annealed Neural Network
Verner, Robert ; Baruník, Jozef (advisor) ; Vošvrda, Miloslav (referee)
The presented master thesis is focused on the stock markets returns analysis using a new type of neural network. First chapter of the thesis describes the underlying theory of the financial time series prediction, Efficient Market Hypothesis and conventional forecasting models. Following part illustrates biological framework, basic principles, functioning of neural networks, their architecture and several well-known learning algorithms such as Gradient descent, Levenberg-Marquardt algorithm or Conjugate gradient. It also mentions certain disadvantages which influence the performance and effectiveness of neural networks. Third chapter is devoted to two applied metaheuristic techniques, i.e. genetic algorithms and simulated annealing that were integrated into neural networks framework to eliminate above mentioned drawbacks. Next chapter describes details of presented hybrid network, whereas the last section is aimed at evaluation of overall results of all models. It shows that on the examined sample hybrid network clearly outperformed standard techniques as well as ordinary neural networks and in most cases achieved the least mean squared error among all explored methods. Keywords: stock returns analysis, neural networks, genetic algorithms, simulated annealing, hybrid networks JEL classification:...
|
|
|
Application of quantile autoregressive models in minimum Value at Risk and Conditional Value at Risk hedging
Svatoň, Michal ; Baruník, Jozef (advisor) ; Vošvrda, Miloslav (referee)
Futures contracts represent a suitable instrument for hedging. One conse- quence of their standardized nature is the presence of basis risk. In order to mitigate it an agent might aim to minimize Value at Risk or Expected Shortfall. Among numerous approaches to their modelling, CAViaR models which build upon quantile regression are appealing due to the limited set of assumptions and decent empirical performance. We propose alternative specifications for CAViaR model - power and exponential CAViaR, and an alternative, flexible way of computing Expected Shortfall within CAViaR framework - Implied Expectile Level. Empirical analysis suggests that ex- ponential CAViaR yields competitive results both in Value at Risk and Ex- pected Shortfall modelling and in subsequent Value at Risk and Expected Shortfall hedging. 1
|
|
|
Volatility modeling : evidence from CEE stock markets
Brabcová, Eva ; Bubák, Vít (referee) ; Baruník, Jozef (advisor)
The thesis applies newly developed heterogenous autoregressive model of realized volatility on high frequency data of three stock market indices: Prague, Budapest and Warsaw with the aim to capture behavior of three different market participants and to quantify their role in forecasting daily realized volatility. Also, the presence of jumps in volatility is investigated and the predictive power assessed. In addition, wavelet analysis is used to detect periods and frequencies of comovements between the three indices. The main contribution of the thesis lies especially in its primary empirical analysis conducted in CEE region. The estimation results indicate that future realized volatility is determined very similarly in all markets with an insignificant impact of participants trading on monthly basis. Moreover, occurrence of a jump proves to be of a high relevance when predicting future volatility. Moreover, wavelet analysis indicates a strong degree of comovement at a frequency of few months across the whole period examined.
|
|
|
Value-at-risk based extreme value theory method and copulas : empirical evidence from Central Europe
Avdulaj, Krenar ; Seidler, Jakub (referee) ; Baruník, Jozef (advisor)
Assessing the extreme events is crucial in financial risk management. All risk managers and and financial institutions want to know the risk of their portfolio under rare events scenarios. We illustrate a multivariate Monte Carlo and semi-parametric method to estimate Value-at-Risk (VaR) for a portfolio of stock exchange indexes in Central Europe. It is a method that uses the non-parametric empirical distribution to capture the small risks and the parametric Extreme Value theory to capture large risks. We compare this method with historical simulation and variance-covariance method under low and high volatility samples of data. In general historical simulation method over estimates the VaR for extreme events, while variance-covariance underestimates it. The method that we illustrate gives a result in between because it considers historical performance of the stocks and also corrects for the heavy tails of the distribution. We conclude that the estimate method that we illustrate here is useful in estimating VaR for extreme events, especially for high volatility times.
|
|
|
Alternative field curve modelling approach : regional models
Šopov, Boril ; Seidler, Jakub (advisor) ; Baruník, Jozef (referee)
In this thesis, we focus on thorough yield curve modelling. We build on extended classical Nelson-Siegel model, which we further develop to accommodate unobserved regional common factors and principal components. We centre our discussion on central European currencies' yield curves: CZK, HUF, PLN and SKK. We propose two novel models to capture regional dynamics; one based purely on state space formulation and the other relying also on principal components of the regional yield curves. Moreover, we supplement the models with two application examples in risk management and structural break detection. The main contribution of this thesis is a creation of a complete framework that enables us to analyse yield curves, to design risk scenarios and to detect structural breaks of various types.
|
|
|
Alternative yield curve modelling approach : regional models
Šopov, Boril ; Seidler, Jakub (advisor) ; Baruník, Jozef (referee)
In this thesis, we focus on thorough yield curve modelling. We build on extended classical Nelson-Siegel model, which we further develop to accommodate unobserved regional common factors and principal components. We centre our discussion on central European currencies' yield curves: CZK, HUF, PLN and SKK. We propose two novel models to capture regional dynamics; one based purely on state space formulation and the other relying also on principal components of the regional yield curves. Moreover, we supplement the models with two application examples in risk management and structural break detection. The main contribution of this thesis is a creation of a complete framework that enables us to analyse yield curves, to design risk scenarios and to detect structural breaks of various types.
|
|
|
Portfolio selectio : clustering algorithm approach
Jenček, Petr ; Pečená, Magda (referee) ; Baruník, Jozef (advisor)
Prices of assets (stocks, commodities etc.) are dependent on many economic factors. These factors may be explicitly known but most of them are hidden. This dependency causes that price of an asset influences prices of another assets which makes it quite complicated to select optimal portfolio. Portfolio management is usually based on various mathematic models in conjunction with Value-at-Risk model. The aim of this thesis is to provide an alternative approach for optimal portfolio selection with mutual assets' prices correlation consideration using cluster analysis.
|
|
|
Causal relationship between Uncertainty and Crude Oil Prices: A Quantile Regression approach
Ruiz Vargas, Andrés Mauricio ; Baruník, Jozef (advisor) ; Krištoufek, Ladislav (referee)
This work considers the causal relationship between the news-based uncertainty measures and WTI crude oil price within the quantile causality framework by using daily data for a period from January 4, 2000, to November 14, 2016. We find that the Granger non-causality test in quantiles between crude oil returns and the news-based uncertainty measures uncover the causal relationship over different levels of conditional quantiles of the crude oil returns. In particular, there exists a strong causal relationship in the tails of the crude oil returns distribution. Powered by TCPDF (www.tcpdf.org)
|
|
|
Modeling multivariate volatility using wavelet-based realized covariance estimator
Baruník, Jozef ; Vácha, Lukáš
Abstract. Study of the covariation have become one of the most active and successful areas of research in the time series econometrics and economic forecasting during the recent decades. Our work brings complete theory for the realized covariation estimation generalizing current knowledge and bringing the estimation to the time-frequency domain for the first time. The results generalize the popular realized volatility framework by bringing the robustness to noise as well jumps and ability to measure the realized covariance not only in time but also in frequency domain. Noticeable contribution is brought also by the application of the presented theory. Our time-frequency estimators bring not only more efficient estimates, but decomposes the realized covariation into arbitrarily chosen investment horizons. Results thus bring better understanding of the dynamics of dependence between the stock markets.
|
|
| |
guest ::
RSS feed.