Churn Prediction Model
Machine Learning · Tirthal Kothari
An interpretable decision tree that connects churn signals with proposed retention actions.
The challenge
This machine-learning project used a subscription dataset with tenure, charges, support calls, and contract type to model customer churn and understand its contributing factors.
What I did
I built a decision-tree classifier, tuned depth and minimum leaf size to limit overfitting, and interpreted the tree’s splits to develop proposed retention actions.
The deliverable
An interpretable churn classifier with proposed retention actions based on its decision rules.
Tools & methods
- Python
- scikit-learn
- Decision Trees
- LabelEncoder
Project context
Machine-learning project. Deliverable: a decision-tree classifier with tuned depth and leaf size, translated into proposed retention actions.