ENHANCING CUSTOMER EXPERIENCE
- Using Customer Churn Prediction and Prevention through Recommendation Engines.
PROBLEM
- Customer churn is a big concern for telecom service providers.
- Full Cost of Customer Churn = Lost Revenue + Marketing Costs.
SOLUTION
- Ability to predict customers at high risk of churning while there is still time to retain them represents huge additional potential revenue.
FEATURES
- Predicting in advance which customers are going to churn.
- Know which marketing actions will have the greatest retention impact.
ANALYSIS
- Accuracy - 94.4 %
- Precision - 0.92 (tp/tp+fp)
- Recall - 0.68 (tp/tp+fn)
USAGE
- ipython notebook file has all the code. Dependencies must be installed in advance. It is preferred to install Anaconda2 on your system before running the code.
- Data Source file is churn.csv
- Output Data files are binary_churn.csv (with binary analysis) and probability_churn.csv (with probability analysis)