storing RMSE values in Proc Entropy

[Serpil]

  Hi,

  I am perfoming a simulation over proc entropy, the basic model is

  proc entropy
data=one GME outest=parm1;

  model y = x1 x2;

run;

 
and my intention is to store “root mean squared errors(RMSE)” to work library for each
run. The file ‘parm1’ does not contain RMSE values. Although it is done easily
for classical regression model, I have yet to find out for proc entropy. 

Regards,

Serpil

[Ksharp]

ods trace on ;

proc entropy;

.............

run;

ods trace off;

Check the name of table(XXXX) for  “root mean squared errors(RMSE)”

then run

ods output XXXX=want;

proc entropy;

[Serpil]

Dear Ksharp, thanks for the tips, but I failed to apply. Below is the entire codes. I'd like to RMS's to appear in the work library. The below case has 10 replications, this means that I will have 10 RMS values for each replication.

Thanks a lot...

Serpil

data one;   

call streaminit(156789);   

do by = 1 to 10;      

do x2 = 1 to 30;

         x1 = 30 * ranuni( 512);

         y = x1 + 2*x2 + rand('chisquare',5);         

output;      

end;   

end; run;

proc entropy data=one GME outest=parm1;   

model y = x1 x2;  

by; run;

[stat_sas · Accepted solution]

Hi,

Just need to make few modifications in the above code to make it work.

Thanks,

data one;   

call streaminit(156789);   

do by = 1 to 10;     

do x2 = 1 to 30;

         x1 = 30 * ranuni( 512);

         y = x1 + 2*x2 + rand('chisquare',5);         

output;      

end;   

end; run;

ods output ResidSummary=ResidSummary;

proc entropy data=one GME outest=parm1;   

model y = x1 x2;  

by by;

run;

proc print data=ResidSummary; /* you will see RMSE for all of 10 replications */

run;

[Serpil]

it worked out. Many thanks...