A comprehensive MLOps solution for predicting which bank clients are most likely to subscribe to term deposits, enabling more effective marketing campaigns.
What It Does
This project helps financial institutions optimize their marketing efforts by identifying prospects with the highest subscription probability. The solution implements a complete machine learning workflow from data processing to model training and evaluation.
How It Works
- Processes banking campaign data with features like customer demographics, past interactions, and economic indicators
- Handles imbalanced data with appropriate sampling techniques
- Uses XGBoost for classification with feature importance analysis
- Evaluates models on metrics optimized for unbalanced classification
- Deploys the model with a ZenML pipeline for continuous training and inference