Application of deep reinforcement learning on automated stock trading
Abstract
How to make right decisions in stock trading is a vital and challenging task for investors. Since deep reinforcement learning (DRL) has outperformed human beings in many fields such as playing Atari Games, can a DRL agent automatically make trading decisions and achieve long-term stable profits? In this paper, we try to solve this challenge by applying Deep Q-network (DQN) and Deep Recurrent Q-network (DRQN) in stock trading and try to build an end-to-end daily stock trading system which can decide to buy or to sell automatically at each trading day. The S...P500 ETF is selected as our trading asset and its daily trading data are used as the state of the trading environment. The agent's performance is evaluated by comparing with benchmarks of Buy and Hold (BH) and Random action-selected DQN trader. Experiment results show that our DQN trader outperforms the two benchmarks and DRQN trader is even better than DQN trader mainly because the recurrence framework can discover and exploit profitable patterns hidden in time-related sequence.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
DRL has been recognized as an effective approach in quantitative finance.
“As deep reinforcement learning (DRL) has been recognized as an effective approach in quantitative finance, getting hands-on experiences is attractive to beginners.”
From FinRL: A Deep Reinforcement Learning Library for Automated Stock Trading in Quantitative Finance (2020) · cites this paper as [8]
The citation lists haven’t been fetched yet
References 13 · Cited by 78
- Authors
- Chen, L., & Gao, Q.
- Affiliations
- Beihang University (CN)
- ORCID
- Venue
- 2019 IEEE 10th International Conference on Software Engineering and Service Science (ICSESS)
- Publisher
- IEEE
- Published
- 2019-10
- Pages
- 29-33
- OpenAlex
- W3012333258
- Cited by
- 78
- References
- 13
- Citations by year
- 2026: 3 · 2025: 8 · 2024: 21 · 2023: 19 · 2022: 13 · 2021: 11 · 2020: 3
- FWCI
- 9.75
- Type
- conference paper
- Language
- en
- Open access
- closed
- Licence
- Locations
- IEEE International Conference on Software Engineering and Service Science (ICSESS) (publishedVersion)
- Topics
- Social Sciences › Decision Sciences › Management Science and Operations Research › Stock Market Forecasting Methods
- Social Sciences › Economics, Econometrics and Finance › Finance › Financial Markets and Investment Strategies
- Social Sciences › Business, Management and Accounting › Accounting › Financial Analysis and Corporate Governance
- Keywords
- automated stock trading, deep reinforcement learning, deep Q-network, deep recurrent Q-network, S&P 500 ETF, buy and sell decisions, recurrent neural networks
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
