Event-Driven Market Prediction with Deep Learning

Market prices for stocks, equities, derivatives, or commodities are highly influenced by economic events. For example, an earnings announcement of a company is known to affect its stock price. Other events include acquisitions, court cases, new product releases, or changes of regulations and laws. Studying the impact of these events is valuable for both traders and risk managers.

This white paper introduces deep learning as a method to perform such event studies. A generic representation of these events is developed, which is then used as an input for a deep neural network. This is trained using historical data, fast and efficiently on CPUs or GPUs. The learner can then be used to study future events and predict their effects on the market.