topic Re: Output sorted list from Scoring Model in SAS Data Science

Output sorted list from Scoring Model

LB3 — Fri, 09 Nov 2018 20:38:08 GMT

I am following the example in this tutorial but using a different data set from a CSV file which contains a list of customers with several categorical input variables along with a binary target variable (1,0) of those that had converted:

support.sas.com/documentation/cdl/en/emgsj/67981/HTML/default/viewer.htm#n0o0v7spwdeoqcn11vnwnuc8x91l.htm

However, when I run the final SAS Code, I get an output of the entire list when what I expected from the tutorial was a list of potential customer conversions. I'm not sure settings might need to be changed because I didn't use the Wizard as indicated on this page:

support.sas.com/documentation/cdl/en/emgsj/67981/HTML/default/viewer.htm#n0uuvzj5du0ohon13mr6xboyf945.htm

but rather simply imported another CSV file and changed the Role to Score to take the place of the sample DONOR_SCORE_DATA with all of the customers who had not yet converted. Any ideas on what could have gone wrong? Let me know if you need any further clarification.

Re: Output sorted list from Scoring Model

Reeza — Fri, 09 Nov 2018 20:42:02 GMT

I would expect to get all data scored and then you'd filter the list based on what is scored as a high potential conversions.

Re: Output sorted list from Scoring Model

LB3 — Tue, 13 Nov 2018 18:43:40 GMT

That's what I'm missing. The SAS Code looks something like this, but obviously it doesn't give me any score. 'Conversion' is my target variable, but that field doesn't exist in the dataset to be scored since they are all Conversion = 0. When I try to add variables such as EM_EVENTPROBABILITY or EM_CLASSTARGET I get an error message. Not sure what variable to use to show what the high potential conversions should be.

proc sort data= EMWS1.Score_SCORE out= bestlist;
by descending State_Code;
run;

proc print data= bestlist;
var TIN SalesIndicator Footprint Tenure State_Code;
run;

SAS Code nodeDataset to be scored