By conducting an analysis on existing customer base of their demographics, service status and attrition records, you are required to analyse the data to provide insights of the churn population and develop/evaluate predictive models for customer retention purpose. Data file named customerChurn_final has been attached.
Following tasks to be performed while preparing the report:
1. Breakdown the characteristics of churned, not-churned customers and loyal customers. Conduct descriptive analysis based on the customer data and construct customer profiles for each customer group. For this, you may compare variables of churned, not-churned and loyal customers using descriptive analytics. Loyal customers are a subset of not-churned customers, they are the top decile of not-churned customers based on tenure index(A re-scaled value of tenure).
2. Using SAS Enterprise Miner, develop and evaluate Regression, Decision Tree and Neural Network models for churn.
a) What are the important variables differentiating churned and not-churned customers?
b) What is the overallchurn rate and the churn rate for various demographics and service groups (for example: senior, has_partner, has_phone, etc.)?
c) What are the predictive performances of various models and how they rank against one another? Make a table of all the machine learning metrics (Accuracy, Precision, Recall or Sensitivity, Specificity, F1 Score, ROC, AUC & Lift) of all the models and how do you interpret the results?
d) How do you best interpret the model? Build a confusion matrix, and explain the correctness and accuracy of the model.
3. Provide campaign recommendations based on insights obtained from Tasks 1 and 2.
You are required to prepare a report with answers to above 3 key tasks, please use screenshots which you feel are important for the body. The report should consist of
a) Table of content
b) Introduction to the report
c) Task 1
d) Task 2
e) Task 3
f) Conclusion
g) Recommendations
h) References
i) Appendix
0 comments