π New Project: Predicting Bank Term Deposits with XGBoost
I've been diving deep into #MLZoomcamp and just finished containerizing a marketing prediction model.
Key highlights:
β Handled imbalanced data (88% vs 12%) using AUC as the primary metric.
β Trained an XGBoost model with hyperparameter tuning in a Jupyter Notebook.
β Built a Flask API to serve real-time predictions.
β Containerized the entire environment with Docker for easy deployment.
π GitHub Repo:
Thanks
#mlzoomcamp
Alexey Grigorev
github.com