This is an experimental Reinforcement Learning based Bitcoin trader program.
The utils folder consists of the functions that download data, pickle the data, get state information and decipher the action that the agent might be taking
This is where the data pickles are stored
The logs are where the logs from the Agent are stored. The Models folder is where the saved models are stored
The scripts folder has the implementation.ipynb which runs the Agent and kicks off the MDP and the learning process.