Portfolio Optimization using AI and Financial Algorithms

Project Structure

• ai_algorithm

  • rl
    • personal_env.py: Custom gymnasium environment for PPO to use and get rewards.
    • stock_rl.py: Using PPO RL algorithm to find the best future prediction.
  • ml
    • price_prediction.py: model to predict future behavior of a stock.

• data: Where data collection happens, and where all the necessary data is stored.

• common: Commonly used functions or constants

  • data_collection.py: sets up all the necessary data files and models
  • config.py: Global constants.

• financial_algorithm

  • black_litterman.py: Takes market caps, price, investor beliefs and calculates optimal weights for the portfolio.
  • evaluate_portfolio.py: Backtesting.
  • fama_french.py: Implements Fama French with 5 parameters.

• main.py: Run this file for desired results.