/forex_rl

Automated trading system for foreign exchange trading using Deep Q-Learning with a deep neural network (CNN) and some intelligent trading strategies

Primary LanguageJupyter Notebook

forex_rl

Directory Structure

  • config: default YAML config containing all the parameters that can be tuned
  • models: details of the model runs contained in the following files:
    • <model time stamp>_config.yaml: contains the configuration of the model run
    • <model time stamp>_metrics.yaml: contains the test set metrics
  • reference: some random reference pdfs
  • scripts: random scripts e.g. run analysis on all the model/*metrics.yaml files
  • src: main code, this contains these important files:
    • fx_env.py: defines the trading environment using a subclass of gym.ENV
    • fx_model.py: defines the deep learning model and memory classes
    • helpers.py: defines helper functions
    • indicators.py: defines functions for calculating technical indicators for trading
    • main.py: main script where the env is created, model is trained, etc.
    • optimze_params.py: bayesian optimization script of all the variables found in config/config.yaml