Abstract | ||
---|---|---|
Recent advance of deep recurrent neural network has brought significant improvements for time series prediction of financial market data. In this paper we present the design, implementation and evaluation of an end-to-end deep learning based trading platform that is able to effectively evaluate quantitative models and perform back-testing, paper trading and real trading using real-time market data streams. This trading system contains all the core components including data collection, model training, model selection, strategy execution, order/risk management and system status monitoring along with a central database. The platform is designed to enable automated learning and tuning capabilities for both prediction model and trading strategy in a rapidly changing market environment. We design our deep learning model based on Long Short Term Memory (LSTM) and perform live paper trading to evaluate the performance and reliability of our approach. Paper trading of several futures contracts on our trading platform achieves promising returns for a one-year time span.
|
Year | DOI | Venue |
---|---|---|
2019 | 10.1145/3321408.3322625 | Proceedings of the ACM Turing Celebration Conference - China |
Keywords | Field | DocType |
deep learning, financial market prediction, neural networks, trading system | End-to-end principle,Computer science,Artificial intelligence,Deep learning,Multimedia | Conference |
ISBN | Citations | PageRank |
978-1-4503-7158-2 | 0 | 0.34 |
References | Authors | |
0 | 5 |