This is a personal side project to explore the use of Deep RL in stock trading. As a start, I am currently trying to reproduce the results of the paper - Deep Robust Reinforcement Learning for Practical Algorithmic Trading. After that, I'd like to explore hierarchical ideas to handle portfolio management.
Note: This is still an incomplete and ongoing project. Please feel free to contribute and provide feedbacks. If you are interested in exploring some new ideas together, I'd love to have some collaborators to learn together!
- Establish a codebase to perform experiments - (60-70% complete)
- Reproduce results of the paper
- Improve documentation and code efficiency
- Explore hierarchical ideas
At the moment, I am using data obtained from Polygon's API. I am experimenting mostly with AAPL data in minute interval.
A3C implementation - https://github.com/dgriff777/rl_a3c_pytorch