Recommending subscription plans with machine learning
Company project
BeCode company project with Orange
A company project during my BeCode traineeship: building a system that recommends the most suitable subscription plan for a customer.
What I did
- Ran a thorough exploratory data analysis to understand the customer data.
- Engineered features that uncovered the key drivers behind plan choice.
- Compared and tuned models with PyCaret and XGBoost.
Result
The final model reached an accuracy of 80.9% on unseen data.
Tools
- Python
- scikit-learn
- PyCaret
- XGBoost