Customer Churn Prediction
Project Information
View on GitHub- Category: Predictive Analytics
- Technologies: Scikit-learn XGBoost Flask Plotly
- Project Date: July 2024
- Dataset: Telco Customer Churn
- Accuracy: 87.5% (cross-validated)
Customer Churn Prediction
A machine learning model that predicts customer churn for companies, helping businesses identify at-risk customers and implement retention strategies.
Technical Implementation
- Multiple classification algorithms compared (Logistic Regression, Random Forest, XGBoost)
- Feature importance analysis using SHAP values
- Handling of class imbalance with SMOTE technique
- Hyperparameter optimization using GridSearchCV
Key Features
- Early identification of at-risk customers
- Customer segmentation based on churn probability
- Interactive dashboard for business stakeholders
- Exportable reports with retention recommendations
Video Demonstration
See the Customer-Churn-Prediction Identification system in action
This demonstration shows the complete workflow of the Customer-Churn-Prediction system.