Realised Volatility Forecasting: Machine Learning via Financial Word Embedding
This is an incremental improvement for financial analysts and traders, enhancing volatility forecasting with news data.
The paper tackled the problem of forecasting realized volatility by incorporating news text through an NLP framework, finding that news-based signals improved statistical performance and yielded economically meaningful gains when combined with standard benchmarks.
We examine whether news can improve realised volatility forecasting using a modern yet operationally simple NLP framework. News text is transformed into embedding-based representations, and forecasts are evaluated both as a standalone, news-only model and as a complement to standard realised volatility benchmarks. In out-of-sample tests on a cross-section of stocks, news contains useful predictive information, with stronger effects for stock-related content and during high volatility days. Combining the news-based signal with a leading benchmark yields consistent improvements in statistical performance and economically meaningful gains, while explainability analysis highlights the news themes most relevant for volatility.