Customer churn predictor
Predicting which customers will leave from how they behave, tuned with Bayesian optimisation.
- Built
- 04-2024
A reproducible machine-learning workflow for predicting when a customer is likely to leave a service based on behavioral data.
The project uses statistical feature selection, XGBoost, and Bayesian hyperparameter optimization, with the workflow kept visible in the public repository for inspection.
Related: Applied ML