This project aims to predict the direction of the S&P 500 index using historical price and volume data. It utilizes machine learning techniques, specifically a Random Forest Classifier, to forecast whether the index will increase or decrease in the future.
- Data Collection: Historical price data for the S&P 500 index is fetched from Yahoo Finance API.
- Preprocessing: Irrelevant columns are removed, and a target variable is created based on tomorrow's price compared to today's price.
- Model Training: A Random Forest Classifier model is trained using historical data, with features such as closing price, volume, open, high, and low.
- Backtesting: The trained model is backtested on historical data to evaluate its performance over different time horizons.
- Feature Engineering: Additional features such as rolling averages, ratios, and trend indicators are engineered to enhance model performance.
- Model Evaluation: Precision score and other evaluation metrics are calculated to assess the model's accuracy.
main.py: Contains the main script for data preprocessing, model training, backtesting, and evaluation.requirements.txt: Lists all the required dependencies for running the project.
- Install the required dependencies listed in
requirements.txt. - Run
main.pyto execute the entire pipeline, from data preprocessing to model evaluation.
- Incorporate more advanced machine learning algorithms.
- Optimize model hyperparameters for better performance.
- Explore additional features and data sources for improved prediction accuracy.
This project utilizes the yfinance library for fetching historical stock data and the scikit-learn library for machine learning implementations.