Replies to This Discussion

Permalink Reply by Rakesh Paul on June 17, 2014 at 12:00pm

Thank you to all for all your suggestions. I got lots of valuable feedback.

• ▶ Reply

Permalink Reply by Dr. Rahul Vyas on July 20, 2015 at 10:40pm

Read and try Survival Analysis Modeling for a churn prediction...

• ▶ Reply

Permalink Reply by Sagar Diwakar Uparkar on August 13, 2015 at 10:04am

Hi good evening everyone.
I read this article little late. You do have many variables. You can use customer demographic data, purchase history, severe usage and billing date. Then you can use logistic regression model to do prediction. Since you do have nominal, ordinal and scale data.

• ▶ Reply

Permalink Reply by Dr. Rahul Vyas on March 18, 2016 at 3:00am

Kindly check the following link for right answer: (An example is given for churn prediction in case study) 

http://www.academia.edu/11846437/Statistical_Analysis_Need_of_Today...

• ▶ Reply

Permalink Reply by Richard Boire on February 19, 2018 at 3:10pm

The most important step in churn is defining your target variable. And more importantly, whatever you define as churn must be in agreement with the business team. Once you have agreement, you can then begin to code for this and then build your analytical file with all the independent variables.  But if the business team asks you to define churn, one good way is to look at activity, be it usage or billing. But you need to define a period of time for where there is no usage or billing and which can be defined as being representative of churn..

  

• ▶ Reply

• ‹ Previous
• 1
• 2
• Next ›
• Page