Overview
Built machine learning models using Gradient Boosting and XGBoost for stock market trend forecasting, achieving R² values up to 0.91 through advanced feature engineering and preprocessing.
Approach
- XGBoost and Gradient Boosting ensemble models
- Feature engineering on price, volume, and technical indicators
- Outlier detection and robust preprocessing
- Walk-forward validation to prevent look-ahead bias
Results
| Metric | Value | |--------|-------| | R² score | up to 0.91 | | Models | XGBoost, Gradient Boosting |
Tech Stack
Python XGBoost scikit-learn pandas NumPy Matplotlib