-Sentiment and Volatility in Cryptocurrency Markets: An Empirical Investigation Using GARCH-X, EVT, and Regime-Switching Models

This thesis investigates the role of sentiment in explaining volatility dynamics within cryptocurrency markets. Using data from Bitcoin and other major cryptocurrencies, the study integrates sentiment-driven variables into GARCH-type models, including GARCH-X and EGARCH-X, to capture the influence of investor mood on market fluctuations. Additionally, Extreme Value Theory (EVT) and Markov regime-switching frameworks are employed to analyze tail risks and volatility regime transitions. The empirical findings demonstrate that sentiment indicators significantly improve the predictive performance of volatility models, particularly during periods of market turbulence. The results highlight the importance of behavioral factors in risk assessment and offer new insights into volatility forecasting in digital asset markets.