Project information
Market risk forecasting using sentiment analysis

Due to the increasing online activity of the general public in recent years, market sentiment has gained significant importance. Not only are the news media shaping public opinions about markets, but social networks, online forums and other internet sources also play a significant role. As public behaviour and opinions change, new challenges in research emerge. The main goal of this project is to discover new ways of capturing market sentiment and incorporating it into various models for market risk prediction (VaR and ES forecasting) in order to improve their predictive power. This will be conducted by identifying suitable markets and dependent variables, collecting vast amount of online data, employing various machine learning approaches and backtesting model results.