Chrun Model / Survival Time Analysis

hazard_function
survival_analysis
customer_churn

#1

Hi All,

I want to predict probability of customer churn given some set of characteristics. These characteristics can be time varying across customer life cycle.

Im trying to find - Given the customer is still active now, what is the probability of him churning in next 30 days?

I read about Survival Analysis, Hazard function gives me probability of a customer churn at that point given he has survived till that time. How do I estimate Hazard function, Cox PH gives me only the Hazard Ratio.

Any help would be highly appreciated.

Thanks
Arvind


#2

Arvind,
For your problem when the window of target variable is certain and hence you need to only deal with uncensored data, I don’t think you even need a survival model.Here are a few things you need to keep in mind before using a survival model (http://www.analyticsvidhya.com/blog/2014/04/survival-analysis-model-you/). A simple logistic or CART model can do this job. But still if you wish to get kick started on survival analysis you can have a look at this article http://www.analyticsvidhya.com/blog/2014/04/solving-survival-model/.

Hope this helps,
Tavish