/numerai_machine_learning_tournament

code for numerai tournament training, test, inference e2e pipeline to help hedge fund stregagy

Primary LanguagePythonApache License 2.0Apache-2.0

README

This is a monolith with full modeling lifecycle participating numerai tournament. It contains E2E pipelines for:

  • training scikit-learn compatible gradient boosting models (such as XGBoost, LightGBM and CatBoost) and ensemble.
  • validate models using industry standard metrics such as Sharpe, Corr, Stdev, MDD.
  • using optuna for hyperparameters tuning
  • custermize objective, metrics and postprocessing such as neutralization
  • weekly inference to meet the tournament requeiremt

Model Performance

This codebase was used for several models: ct_balance was on the weekly top 5%/15% (20+ times) performance within 1 yr time frame: https://numer.ai/ct_balance

Info for NumerAI tournament

https://docs.numer.ai/tournament/learn