12-04-2015 10:49 AM - edited 12-04-2015 10:49 AM
I am a beginner in modeling and preparation of data for modeling.
Currently, we prepare the data for modeling churn customers in the TELCO and I have the following problem.
I churn for the period 201505 and to join these data variables for say 6-9 months before the churn rate and it will targer churn = 1.
The problem is with the TARGET = 0, which here should take contracts?
Is such a database only from the period 201411 and variable calculated from 201411 to 201505? (only those which do not churn)?
Should I take the entire database, or for example if I examine the commercial churn (201505) to take the base of such 201411,
which will EOP (End of period) in the coming months (eg 6 months)?
Apologizes for probably a trivial question, but as I wrote above, I am a beginner.
Thank you very much for your help, advice, suggestions on what I should pay attention to.
12-04-2015 11:27 AM
02-18-2016 11:34 AM
where is the paper on this subject? Im interested in doing the same thing and I just want to get an idea on how the data should be structured. For example i want to get propensity to churn in the next 30 days in telco. So i have my customer base this month should i go 12 months back and have data for each month?
02-18-2016 02:22 PM
Here are a couple resources on survival data mining.
(See videos Introduction to Survival Data Mining and New Features in the SAS Enterprise Miner 12.3 Survival Node)
and this tip that gives a link to an example SAS Enterprise Miner template and accompanying documentation for survival data mining:
02-18-2016 02:31 PM
You can search lexjansen.com with the terms: survival analysis with churn from a telecom company
You'll find many papers and hopefully something that works for you.