A growing number of research has found evidence for the presence of excess information about future volatility and returns in financial news that is not present in the time series of relevant assets.
Engaging with researchers from MAN group, teams of OMI researchers are investigating the potential of state-of-the-art machine learning and natural language processing techniques to automatically uncover and exploit publicly available textual information in order to gain a predictive edge over pure time series approaches. Applications vary between projects and span a range of asset classes, including equities, FX and cryptocurrencies. In collaboration with researchers at Harvard, we are also studying the potential exploitation of news sentiment propagation and network effects.
DPhil Student, Engineering, Machine Learning Research Group