06-07-2016 10:12 AM
I am trying to build a dataset for all existing credit cards in our portfolio to track if any of the accounts would go into default whithin 6 or 12 months. does anyone have any ideas about how to collect this dataset? I am using a logistic regression to model account behavior.
06-07-2016 11:23 AM
Are you asking how to import some available data in a sas dataset for further processing ?
If so, you will probably get more feedback if you post in a more general forum (Base SAS programming for instance) and
avoid technical terms (behavioral model, logistic regression, ...) that are not strictly relevant to your problem.
If your problem really concerns importing data in SAS, you should also give some precisions on the data at your disposal
(file type, structure,...)
If I completely miss your point, sorry for the irrelevant advice.
06-07-2016 04:31 PM
To be able to model credit cards going into default in the next 6 or 12 months you first need historical credit card data. You also need a rule to define default if your source data does not already define that.
Say for example you select all open credit cards as at March 2015. You then need to select data for all of the same credit cards between April 2015 and April 2016 and identify any cards that went into default over that period. You then need to repeat this process for different starting months to cater for any changes in default behaviour over time.
Hope this gives you something to start with. It takes a lot of time and effort to bring all of the required data together to build a reliable model. Also the more historical data you have the better.