|
|
MCMC methods for financial time series
Tritová, Hana ; Pawlas, Zbyněk (advisor) ; Komárek, Arnošt (referee)
This thesis focuses on estimating parameters of appropriate model for daily returns using the Markov Chain Monte Carlo method (MCMC) and Bayesian statistics. We describe MCMC methods, such as Gibbs sampling and Metropolis- Hastings algorithm and their basic properties. After that, we introduce different financial models. Particularly we focus on the lognormal autoregressive model. Later we theoretically apply Gibbs sampling to lognormal autoregressive model using principles of Bayesian statistics. Afterwards, we analyze procedu- res, that we used in simulations of posterior distribution using Gibbs sampling. Finally, we present processed output of both simulated and real data analysis.
|
|
|
The fractal dimension and forecasting of financial time series
Kaplan, Robert ; Krištoufek, Ladislav (advisor) ; Džmuráňová, Hana (referee)
In this thesis, we strive to build on the fractal market hypothesis and to develop two methods which aim to reveal whether the fractal dimension, as a property of the short memory, can be applied for forecasting of financial time series. In the first one, we use ten world market indices and repeatedly estimate the fractal dimension by boxcount, Hall-Wood, and Genton estimators on fixed number of returns and make one step ahead forecasts by AR(1) and ARMA(1,1) models; then, we look whether forecast errors from realized returns are lower when the fractal dimension is estimated lower. The second method incorporates only the fractal dimension and studies, if the sign of return persists in next period more likely with lower fractal dimension. The results indicate that the short memory is truly present in the markets and the fractal dimension may be potentially useful for prediction and increased profit for investors. However, the significance of our results is not strong. We recommend more sophisticated methods and models for further research.
|
|
|
Analysis of financial time series with economical news headlines
Kalibán, František ; Petrásek, Jakub (advisor) ; Zichová, Jitka (referee)
This thesis is focused on options of improving the estimate of volatility of the given financial time series by analysing the economical news headlines. Because of very large volume of data and correlation between word occurence in headlines, the Principal Component Analysis is used to reduce the dimension of data space. For the elimination of significantly large skewness of dependent variable and the preservation of its normality a Box-Cox transformation is used. Finally, a linear model is constructed and its robustness is analyzed by cross-validation method. The computations were made by R software.
|
|
|
Multivariate Financial Time Series
Veselý, Daniel ; Cipra, Tomáš (advisor) ; Kopa, Miloš (referee)
In this work we will describe methods for modeling multivariate financial time series. We will concentrate on both modeling expected value by multi- variate Box-Jenkins processes and primarily on modeling conditional corre- lations and volatility. Our main object will be DCC (Dynamic Conditional Correlation) model, estimation of its parameters and some other general- izations. Then we will programme DCC model in statistical software R and apply on real data. In applications we will concentrate on problem of high dimension of financial time series and on modeling conditional correlations data with outliers.
|
|
|
Technical analysis of financial time series
Faltýnková, Anežka ; Petrásek, Jakub (advisor) ; Hurt, Jan (referee)
The thesis studies the problem of inefficiencies in the finan- cial markets. The first section describes the fundamental concepts, such as the efficient market hypothesis and futures contracts. The necessary mathematics is summarized in the second part, which deals with the link between the futures price and the martingale. The nonlinear regression is introduced and the greatest emphasis is placed on the description of the functional linear model with a scalar response. The main part focuses on the application of this theory. Two models are proposed for predicting prices based on their historical changes. The first model is nonlinear and is based on the assumption that the impact of the price change on the prediction process diminishes exponentially with time. The second one is linear and directly estimates the effect of particular changes. Both models are compared in terms of their ability to predict inefficiencies, calculation costs and stability. 1
|
|
|
Robustní testy normality a jejich využití při ověřování slabé formy efektivnosti akciového trhu
Střelec, Luboš
Submitted dissertation is focused on methods of robust normality testing and applications of robust tests in verifying hypothesis of the weak form of efficiency in stock markets. In the dissertation, theory of efficient markets and approaches to verifying the weak form of market efficiency and normality assumption are being discussed. Novel robust testing procedures of testing normality are proposed in this work to overcome shortcomings of classical normality tests in the field of financial data, which are typical with occurrence of remote data points and additional types of deviations from normality. Results of power simulation study of classical and robust tests of normality against several types of alternative distributions, i.e. symmetric heavy-tailed, symmetric light-tailed, asymmetric heavy-tailed, asymmetric light-tailed, selected mixtures of normal distributions and outlier models, are presented. Based on outcome of the power simulation study, selected normality tests were consequently used to verify the weak form of efficiency in stock markets in the Czech Republic, Hungary, Austria, Germany, Slovakia, United States and Japan during years 2000-2009. In addition to selected classical and robust normality tests, Ljung-Box portmanteau test was also used. In conclusion, there is a discussion and comparison of results carried out and future trends of these markets are outlined.
|
|
|
CRUDE OIL PREDICTION FOR COMPANIES IN ENERGY DEMANDING PRODUCTION
Vícha, Tomáš ; Dohnal, Mirko (advisor)
The dissertation deals with prediction of crude oil price and is tailor-made for such companies which are heavily crude oil related. The main dissertation target is to make sure that such companies can get ready for price changes and safeguard themselves against negative consequences. Crude oil prices are the main factor which affects prices of such final products as petrol. It is a well known fact that quantitative predictions are not reliable and all those who are forced to real on such vague data set for their decision-making are reluctant to use them. That’s how we would like to have at least the correct trend information. The dissertation introduces some concepts originally developed within artificial intelligence theory for the crude oil predictions. Specifically common sense algorithms and qualitative interpretation of some aspects of theory of chaos are the main contribution towards expanding of available prediction tools described by the dissertation. A systematic analysis of a sequence of qualitative solutions is the key part of the dissertation.
|
|
|
Modeling and Forecasting Volatility of Financial Time Series of Exchange Rates
Žižka, David ; Arltová, Markéta (advisor) ; Malá, Ivana (referee) ; Vošvrda, Miloslav (referee)
The thesis focuses on modelling and forecasting the exchange rate time series volatility. The basic approach used for the conditional variance modelling are class (G)ARCH models and their variations. Modelling of the conditional mean is based on the use of AR autoregressive models. Due to the breach of one of the basic assumption of the models (normality assumption), an important part of the work is a detailed analysis of unconditional distribution of returns enabling the selection of a suitable distributional assumption of error terms of (G)ARCH models. The use of leptokurtic distribution assumption leads to a major improvement of volatility forecasting compared to normal distribution. In regard to this fact, the often applied GED and the Student's t distributions represent the key-stones of this work. In addition, the less known distributions are applied in the work, e.g. the Johnson's SU and the normal Inverse Gaussian Distribution. To model volatility, a great number of linear and non-linear models have been tested. Linear models are represented by ARCH, GARCH, GARCH in mean, integrated GARCH, fractionally integrated GARCH and HYGARCH. In the event of the presence of the leverage effect, non-linear EGARCH, GJR-GARCH, APARCH and FIEGARCH models are applied. Using suitable models according to the selected criteria, volatility forecasts are made with different long-term and short-term forecasting horizons. Outcomes of traditional approaches using parametric models (G)ARCH are compared with semi-parametric neural networks based concepts that are widely applicable in clustering and also in time series prediction problems. In conclusion, a description is given of the coincident and different properties of the analyzed exchange rate time series. The author further summarized the models that provide the best forecasts of volatility behaviour of the selected time series, including recommendations for their modelling. Such models can be further used to measure market risk rate by the Value at Risk method or in future price estimating where future volatility is inevitable prerequisite for the interval forecasts.
|
|
|
Weather influence on speculation on stock markets
Horáček, Jan ; Pánková, Václava (advisor) ; Křepelová, Marika (referee)
Topic of this master thesis is to examine whether weather related mood changes are in correlation with price of stocks. Thesis focuses on middle Europe stock market indexes PX, SAX, ATX and DAX. Research is based on relationship between daily cloud cover and development of the indexes form 1995 to 2012. It also focuses on comparison of several different models, especially models of seemingly unrelated regressions. It shows that indexes PX and ATX are significantly negatively correlated with local cloud cover. Use of seemingly unrelated regressions offers slightly better results. The relation between cloud cover and stock indexes is not strong enough to be used for weather based speculations
|
|
|
Application of R/S Analysis at Financial Markets
Vilhanová, Vanda ; Trešl, Jiří (advisor) ; Kodera, Jan (referee)
The aim of this graduation thesis is the descriptiton of R/S analysis and it's aplication on chosen time series of share prices and exchange rates. Some main models of financial time series will be mentioned in the second chapter. There will described basic linear models of stationary and non stationary time series and models of volatility. Then we will focus on the main theme of this thesis, R/S analysis. The algorithm of R/S analysis and the interpretation of the Hurst exponent will be described in the forth chapter. In the fifth chapter, the R/S analysis will by applied on real data sets. There will be two data sest of share prices of Telefónica O2 and Philip Morris and two data sets of exchange rates CZK/EUR and CZK/USD. The results will be interpreted and compared.
|
guest ::
