Deep Reinforcement Learning for Portfolio Management
The main focus of this research paper is to study Deep Reinforcement Learning and replicate trading strategies based on Convolutional Neural Network. Deterministic Policy Gradient and Deep Deterministic Policy Gradient algorithms are selected to update our Reinforcement Learning Portfolio Manager. The time-series analysis and cross-sectional analysis are also considered in the Convolution Neural Network Construction
MCPG file contains all the Monte Carlo Policy gradient methods to update the corresponding actor
DDPG file contains all the Deep Deterministic Policy gradient methods and replay buffer to update the corresponding actor and critic The design of the agent is inspired by "What is the value of the cross-sectional approach to deep reinforcement learning?"
Env file contains the multi-stock environment which designed in "A Deep Reinforcement Learning Framework for the Financial Portfolio Management Problem"