survival analysis SAS EM - Scoring issues (on full expanded data or transactional data)

[harsh0404]

I am doing survival analysis on donor’s dataset. I am stuck on scoring part. 

I am using transactional dataset and choosing full expanded data option in miner, so the only time dependent covariate is money the donors give every time (dataset I am using in figure1 ). Now from the survival analysis pdf ( http://documentation.sas.com/?docsetId=emex&docsetTarget=n1gyqhds6xjtgcn1qam5r83or9re.htm&docsetVers... ), I am not getting idea on how I should prepare data. The documentation says do it by month or week and the time changing covariate is number of complaints (aggregation of number of complaints by week or month). But in my dataset I am just doing it by transaction, so I am not preparing dataset by months or week transactions of donors. Am I right or wrong here?

In this very approach. I can’t score my dataset, because the score node just gives multiple values for each id (figure2) . Now in the documentation, it tells me the way to score one row per customer data, which is by creating _T_ variable. But for the transactional dataset I already created _t_ variable as described in documentation. Also I don’t see the option to click on flow code (its disabled).

figure 1 

figure 2 

[MikeStockstill]

Hello harsh0404-

Thank you for your interest.

Please open a track with Technical Support so that we can take a closer look.  After you receive an auto-reply email message with a tracking number from the form, send a model package and a subset of the data that you want to score.

We can then advise.

Have a great day.