|
|
News Feed Classifications to Improve Volatility Predictions
Pogodina, Ksenia ; Šopov, Boril (advisor) ; Červinka, Michal (referee)
This thesis analyzes various text classification techniques in order to assess whether the knowledge of published news articles about selected companies can improve its' stock return volatility modelling and forecasting. We examine the content of the textual news releases and derive the news sentiment (po larity and strength) employing three different approaches: supervised machine learning Naive Bayes algorithm, lexicon-based as a representative of linguistic approach and hybrid Naive Bayes. In hybrid Naive Bayes we consider only the words contained in the specific lexicon rather than whole set of words from the article. For the lexicon-based approach we used independently two lexicons one with binary another with multiclass labels. The training set for the Naive Bayes was labeled by the author. When comparing the classifiers from the machine learning approach we can conclude that all of them performed similarly with a slight advantage of the hybrid Naive Bayes combined with multiclass lexicon. The resulting quantitative data in form of sentiment scores will be then incorpo rated into GARCH volatility modelling. The findings suggest that information contained in news feeds does bring an additional explanatory power to tradi tional GARCH model and is able to improve it's forecast. On the...
|
|
|
Alternative Formulation of Pay-as-clear Auction in Electricity Markets
Aussel, D. ; Červinka, Michal ; Henrion, R. ; Pištěk, Miroslav
In widely used formulation of pay-as-clear electricity market the clearing price is given by the Lagrange multiplier of the demand sat- isfaction constraint in the problem of the Independent System Operator (ISO). Following this idea, one may usually calculate the market clearing\nprice analytically even for problems of higher dimensions. However, the economic interpretation of such a market setting is in question, since the minimized criterion does not correspond neither to the cost of production nor to the overall payment of consumers. This observation motivated us\nto propose an alternative clearing mechanism where the total payment of consumers is explicitly minimized. We show existence and uniqueness of the clearing price in such a setting.
|
|
|
Game Theory Approach to Hostile Takeovers
Tuček, Lukáš ; Červinka, Michal (advisor) ; Čornanič, Aleš (referee)
This bachelor thesis investigates the field of hostile takeovers predominantly from the perspective of the management of the target company. Particular emphasis is placed on review of the defense strategies against hostile takeovers and ways in which they might be abused by the management. This thesis attempts to formulate a game-theoretic model describing the process of a hostile takeover as an extensive-form game with perfect in- formation. Payoff functions for the game further in the thesis computed as Nash equilibria of bargaining problem with the current state of the game as a base for the utility gains. We then briefly discuss the implied relationship among the raider, management and the shareholders. JEL Classification C7; C72; G34 Keywords hostile takeover; game in extensive form; defense strategy
|
|
|
Impact of Terrorism on Stock Markets
Koščo, Marek ; Červinka, Michal (advisor) ; Nevrla, Matěj (referee)
Terrorism generally induces negative mood in the society. Financial markets performance exhibits the contingency on the mood of their trading parti- cipants. The thesis enhances the understanding of this interrelated entities by analysing the situation from 2000 to 2015 at the 20 world largest mar- kets. Their composite indices are put under scrutiny employing a multifactor model, a difference equation and a logit model. The impact is confirmed and further discussed, while the logit model provides a simple framework for forecasting index returns just after an attack with more than 25 casualties. Keywords global financial markets, terrorism, multifactor model, difference equation, logit model
|
|
|
Analysis of nonprice competitiveness: Czech Republic
Petrů, Vojtěch ; Semerák, Vilém (advisor) ; Červinka, Michal (referee)
In general discussion, the concept of country's international competitiveness is frequently used for analysing its macroeconomic performance and is usually associated with price competitiveness, often measured by real effective exchange rate. In spite of its usefulness, this indicator has several drawbacks, stemming from its strong assumptions. Using highly disaggregated data from UN Comtrade trade database, a relative export price index which accounts for non-price factors such as changes in the market power and taste or quality is presented and utilized for examination of changes in Czech republic's exports competitiveness between 1998 and 2015 in various product sections and geographical regions. The results show that the appreciation of Czech real effective exchange rate, which occured between 1998 and 2008, was transfered into higher export prices in a very muted form and the improvements in quality of Czech exports were more than capable of compensating for this increase. Major differences in quality growth between individual product sections and exports into different locations are detected. The highest improvements are achieved in products of Machinery, Plastics and rubber, Chemical products, but also Metals.
|
|
|
Least Absolute Deviations
Pacák, Daniel ; Víšek, Jan Ámos (advisor) ; Červinka, Michal (referee)
This is a theoretical study of the Least Absolute Deviations (LAD) fits. In the first part, fundamental mathematical properties of LAD fits are established. Computational aspects of LAD fits are shown and the Barrodale-Roberts Al- gorithm for finding LAD fits is presented. In the second part, the statistical properties of LAD estimator are discussed in the concept of linear regression. It is shown that LAD estimator is a maximum likelihood estimator if the er- ror variables follow Laplace distribution. We state theorems establishing strong consistency and asymptotic normality of LAD estimator and we discuss the bias of LAD estimator. In the last section, we present the results of numerical experi- ments where we numerically showed consistency of LAD estimator, discussed its behaviour under different distributions of error variables with comparison to the Ordinary Least Squares (OLS) estimator. Lastly, we looked at the behaviour of LAD and OLS estimators in the presence of corrupted observations. 1
|
|
|
Bankruptcy prediction models in the Czech economy: New specification using Bayesian model averaging and logistic regression on the latest data
Kolísko, Jiří ; Princ, Michael (advisor) ; Červinka, Michal (referee)
The main objective of our research was to develop a new bankruptcy prediction model for the Czech economy. For that purpose we used the logistic regression and 150,000 financial statements collected for the 2002-2016 period. We defined 41 explanatory variables (25 financial ratios and 16 dummy variables) and used Bayesian model averaging to select the best set of explanatory variables. The resulting model has been estimated for three prediction horizons: one, two, and three years before bankruptcy, so that we could assess the changes in the importance of explanatory variables and models' prediction accuracy. To deal with high skew in our dataset due to small number of bankrupt firms, we applied over- and under- sampling methods on the train sample (80% of data). These methods proved to enhance our classifier's accuracy for all specifications and periods. The accuracy of our models has been evaluated by Receiver operating characteristics curves, Sensitivity-Specificity curves, and Precision-Recall curves. In comparison with models examined on similar data, our model performed very well. In addition, we have selected the most powerful predictors for short- and long-term horizons, which is potentially of high relevance for practice. JEL Classification C11, C51, C53, G33, M21 Keywords Bankruptcy...
|
|
| |
|
|
Least Absolute Shrinkage and Selection Operator Method
Vorlíčková, Jana ; Červinka, Michal (advisor) ; Rusnák, Marek (referee)
The main intention of the thesis is to present several types of penalization techniques and to apply them in economic analyses. We focus on penalized least squares, with a main topic being the lasso. The penalization methods are commonly employed to data sets with a relatively large number of the variables as compared to the sample size. These methods simplify the model by shrinkage of the estimates of the coefficient of the irrelevant variables to wards zero or they put these estimates equal to zero, i.e. they produce a sparse solution. Namely, we present the following methods: ridge regres sion, best subset selection problem, lasso and elastic net. We discuss several applications of the lasso in the current economic and finance research and hence present the lasso in more details. In the practical part of the thesis, we analyze a real economic data using the elastic net, the ridge regression, the lasso and the ordinary least squares method. We use the mean squared error as the measure of performance of the respective method. The penal ized least squares methods surpass the ordinary least squares method, with the elastic net being the best performing method. Keywords penalized least squares, lasso, elastic net, ridge regression, penalization tech niques in economics 1
|
|
|
Player Skill Rating for Games with Random Matchmaking
Hubík, Jan ; Červinka, Michal (advisor) ; Matoušek, Jindřich (referee)
Traditional skill ratings are not suitable for new types of games. We developed a general skill rating framework for games which do not discriminate players based on their skill. This class of games is widely present in the world. We use Bayesian statistics to convert aggregate data about the player's performance to a percentile rank describing his skill. The system is applicable to both single-player and multiplayer games with binary and non-binary endings. The rating formulas do not contain any arbitrary constants. We have tested the system in simulations and on real game data, and we outline its possible applications.
|
guest ::
