PROC LIFEREG: How do I compute predicted value for out of sample observations?

PaulN — Fri, 07 Nov 2025 15:48:46 GMT

Here's my SAS code:

ods trace on;
ods graphics on/imagemap=on;

/* ods output/classlevels corrB covB modelinfo nobs fitstatistics FitStatisticsUL ConvergenceStatus
Type3Analysis ParameterEstimates ProbPlot; */
proc lifereg data=health.E_Ltestvar1_10_70_firms plots=probplot 
		outest=health.estimates covout;
	a: model wks_surv*censoring(0)=ln_AGE complex mmso preneg prepack RSA clntbase 
		DIP mgtresign fraud / distribution=weibull alpha=0.05 corrb covb;
	output out=health.residuals sres=s_residuals; 
	output out=health.xbeta Xbeta=xbeta;
	output out=health.new cdf=probe;
	probplot ppout ppos=mkm npintervals=simul;
	inset nobs="Sample Size" left="Number of left-censored observations" 
		right="number of right-censored observations" 
		interval="Number of interval-censored observations" 
		nmiss="Number of observations with missing values" dist="Distribution" shape 
		confidence="Confidence coefficient for all confidence intervals";
	probplot ppos=km npintervals=point;
	inset;
run;

ods trace off;
ods graphics off;

Here are my SAS datasets. I'm having trouble figuring out how to write the code. I want to take the parameter estimates from the SAS code above and compute the predicted values for the observations (3) in the outsample.sas7bdat file. Then, I want to compare the predicted values to the actual values. The actual values are in the outsample.sas7bdat file and the variable name is wks_surv. How do I do this? Any help is appreciated.

topic PROC LIFEREG: How do I compute predicted value for out of sample observations? in Statistical Procedures

PROC LIFEREG: How do I compute predicted value for out of sample observations?

Re: PROC LIFEREG: How do I compute predicted value for out of sample observations?