an automated trading bot using reinforcement learning
DDQN_PER.py = double deep Q learning with prioratized experience replay
DQN_PER.py = deep Q learning with prioratized experienced replay
dqn_agent.py = deep Q learning with experience replay
model.py = Architecture of the neural network
stock_trader_PER_trend.py = stock trader with DQN and prioritized experienced replay
stock_trader_trend_DDQN_PER.py = stock trader with DDQN and prioritized experienced replay
stock_trader_with_trend.py = stock trader with DQN and experienced replay
Currently, I am working on stock_trader_DDQN_PER.py (stock trader using deep double Q learning with prioratized experience replay) file. So, the results are shown using deep double Q learning with prioratized experience replay.
Result in training set. Training set is 2018 walmert stock market.
Result in test set. Test set is 2019 walmert stock market.
- Run any of the stock trader named file according to your chosen algorithm, module will be automatically imported.
Do not change the folder structure. you can also observ the average Q value and average reward at each episode in tensorboard and those tensorboard files will be stored at runs folder.