RLPortfolio Documentation
RLPortfolio is a Python package which provides several features to implement, train and test reinforcement learning agents that optimize a financial portfolio:
A training simulation environment that implements the state-of-the-art mathematical formulation commonly used in the research field.
Two policy gradient training algorithms that are specifically built to solve the portfolio optimization task.
Four cutting-edge deep neural networks implemented in PyTorch that can be used as the agent policy.
Note
This project is mainly intended for academic purposes. Therefore, be careful if using RLPortfolio to trade real money and consult a professional before investing, if possible.
Environment
Algorithm
Package Reference