I built an XGBoost classifier that identifies real exoplanets from NASA Kepler data. Instead of using diagnostic flags like other competitors (which inflates accuracy to 94%+), I used only raw physics features. Lower score, but honest methodology and actually interpretable results.
Built With
- jupyter
- matplotlib
- pandas
- python
- scikit-learn
- xgboost
Log in or sign up for Devpost to join the conversation.