Deep Order Flow Imbalance: Extracting Alpha at Multiple Horizons from the Limit Order Book
Abstract
We describe how deep learning methods may be applied to forecast stock returns from high frequency order book states. I will review the literature in this area and describe a study where we evaluate return forecasts for several deep learning models for a large subset of symbols traded on the Nasdaq exchange. We investigate whether transformation of the order book states is necessary and we relate the performance of deep learning models for a symbol to its microstructural properties. We also provide some color on hyperparameter sensitivity for the problem of high frequency return forecasting as well as a discussion of the importance of Seq2Seq based architectures for prediction. This is based on joint work with Petter Kolm and Jeremy Turiel.
Bio
Nick is currently head of Execution Research & co-head of Trading in the Multi-Asset Solutions group at AllianceBernstein, where he focusses on automating and improving execution across Equities, Futures and FX. He is also a visiting researcher in Financial Machine Learning at the Courant Institute of Mathematical Sciences at NYU working on problems at the intersection of optimal execution, market microstructure and deep learning. Previously he was a Senior Quant Researcher in the Equity Execution group at Citadel focusing on block trading and market Impact. Prior to that he was at Deutsche Bank involved in the Central Risk Book and Algorithmic Trading. He holds a PhD from Imperial College London and was a Postdoctoral Research fellow at Humboldt Universitaet zu Berlin.