Využití entropie při detekci systémového rizika na finančních trzích

Abstract

This thesis focuses on the application of selected entropy-based measures for detecting systemic risk in financial markets and their use in investment strategies. The main objective is to empirically verify whether Shannon, Tsallis, and Rényi entropy can capture changes in market dynamics and whether these signals can be used to outperform a passive Buy & Hold strategy. The theoretical part defines key concepts of financial markets, investment processes, and risk management, with a focus on volatility, Value at Risk, and Conditional Value at Risk. It also introduces the principles of information entropy and its application in financial economics. The empirical part is based on historical data of selected stock and commodity indices. An algorithm based on alarm rules is used to generate investment signals derived from entropy dynamics, determining the allocation between risky assets and cash. The performance of the strategies is evaluated using standard return and risk measures. The results indicate that entropy-based strategies are capable of responding effectively to changes in market conditions and, in some cases, achieve better performance than a passive approach while maintaining an acceptable level of risk. At the same time, the results show that the performance of the strategies is sensitive to the model parameter settings. The thesis confirms that entropy measures represent a promising tool for analyzing complex financial systems and can contribute to improved risk management and investment performance.