topic Re: SAS procedures for Internal validation using Cross-validation/ Bootstrapping in SAS Programming

SAS procedures for Internal validation using Cross-validation/ Bootstrapping

sms1891 — Tue, 09 May 2023 21:41:49 GMT

Hi SAS experts,

I have a data set with N= 2196 observations, a scoring variable (PIPE_Score values from 1 to 13) and an outcome variable (Outcome, values 0 for No and 1 for Yes). I need help with calculating the sensitivity and specificity for predicting the outcome based on the PIPE_Score values and some kind of internal validation (cross-validation or bootstrapping) to check sensitivity and specificity for predicting the outcome based on the PIPE_Score values using the validation cohort. How can I achieve this? Any help with this is much appreciated.

Thank you so much in advance!

Re: SAS procedures for Internal validation using Cross-validation/ Bootstrapping

ballardw — Tue, 09 May 2023 22:53:50 GMT

If you generated that data set with one of the modeling procedures you should include the code from the modeling procedure.

Then someone more familiar with the procedures can suggest specific options, especially if they may be needed in the model procedure code, to help.

Re: SAS procedures for Internal validation using Cross-validation/ Bootstrapping

sms1891 — Wed, 10 May 2023 15:34:21 GMT

Thank you for the response. I have an excel sheet with the original data. Please see the attached. I do not have any sas code for the question posted because I am not sure how to get it done.

PROC IMPORT OUT= WORK.Derv
            DATAFILE= "\\Desktp\Derv.xlsx" 
            DBMS=EXCEL REPLACE;
     RANGE="Sheet1$"; 
     GETNAMES=YES;
     MIXED=NO;
     SCANTEXT=YES;
     USEDATE=YES;
     SCANTIME=YES;
RUN;

Re: SAS procedures for Internal validation using Cross-validation/ Bootstrapping

ballardw — Wed, 10 May 2023 15:45:32 GMT

When i look at that XLSX I see pictures, nothing that could be "imported".

Is that a "validation" set? Usually validation comes with a training set. Do you have such?

So you are actually asking for help creating the predication and then validating? What kind of rules do you have for the prediction?

Re: SAS procedures for Internal validation using Cross-validation/ Bootstrapping

sms1891 — Wed, 10 May 2023 16:36:41 GMT

The excel sheet had ID variable, Outcome variable and scoring variable PIPE_Score. I want to get sensitivity and specificity calculation for Outcome prediction for PIPE_Score cutoffs of 7, 12 and 15. I also want to generate a bootstrapped cohort and calculate the sensitivity and specificity for Outcome prediction for the same PIPE_Score cutoffs of 7, 12 and 15. I have attached the excel sheet again. Thank you for your help!