Forecasting financial markets with semantic network analysis in the COVID‐19 crisis

This paper uses a new textual data index for predicting stock market data. The is applied to large set of news evaluate the importance one or more general economic-related keywords appearing in text. assesses keywords, based on their frequency use and semantic network position. We apply it Italian press construct indices predict bond returns volatilities recent sample period, including COVID-19 crisis. evidence shows that captures different phases financial time series well. Moreover, results indicate strong predictability data, both volatilities, short long maturities, volatility.