← Back to Projects Telecom 
Customer Churn Prediction with AutoML
Automated ML pipeline for churn prediction using H2O.ai AutoML and Optuna-tuned LightGBM, with Boruta feature selection and business impact calculator.
Overview
Automated ML pipeline for customer churn prediction with business impact quantification.
Architecture
- H2O.ai AutoML for baseline model search
- Optuna-tuned LightGBM for production model
- Boruta and mutual information feature selection
- Business impact calculator translating churn to revenue-at-risk
- Click CLI for pipeline orchestration
Key Features
- Automated model selection and tuning
- Revenue-at-risk business reporting
- Feature importance with multiple methods
- Reproducible pipeline via CLI