Deep Learning for Spatio-Temporal Modeling: Dynamic Traffic Flows and High Frequency Trading
This work addresses predictive modeling challenges in transportation and finance, but appears incremental as it applies standard deep learning techniques to these domains.
The authors developed deep learning architectures for spatio-temporal modeling, applying them to predict traffic flow discontinuities and classify short-term futures market prices based on order book depth.
Deep learning applies hierarchical layers of hidden variables to construct nonlinear high dimensional predictors. Our goal is to develop and train deep learning architectures for spatio-temporal modeling. Training a deep architecture is achieved by stochastic gradient descent (SGD) and drop-out (DO) for parameter regularization with a goal of minimizing out-of-sample predictive mean squared error. To illustrate our methodology, we predict the sharp discontinuities in traffic flow data, and secondly, we develop a classification rule to predict short-term futures market prices as a function of the order book depth. Finally, we conclude with directions for future research.