It's About Time: Discrete Time Survival Analysis Using SAS ® Enterprise Miner™

The new survival analysis algorithm in SAS Enterprise Miner 7.1 provides analysts with an alternate approach to modeling the probability of customer behavior events. Traditional binary classification approaches provide a snapshot view of event propensities, while survival analysis can generate a time based function of event probabilities. This time-based view can help organizations to optimize their customer strategies by gaining a more complete picture of customer event likelihoods. This paper briefly explains the theory of survival analysis and provides an introduction to its implementation in SAS Enterprise Miner. An example illustrates the usage of this analytical algorithm using a customer churn data set.