I have a data set containing 980 paired observations of fish length and age. I would like to use a bootstrapping approach to examine how sample size affects the precision of model parameter estimates.
I have used the following program to successfully create a dataset where 3 fish per cm length class are selected 1000 times from the original dataset:
PROC SURVEYSELECT DATA = AGE OUT = SAMPLE
METHOD = URS
N = 3
SEED = 9876
OUTHITS
REP = 1000;
STRATA TLINT;
RUN;
Next I fit model parameters to each replicate in this dataset:
PROC SORT DATA = SAMPLE;
BY REPLICATE;
RUN;
PROC NLIN DATA = SAMPLE;
PARMS LINF = 550 K = .25 t0 = 0.1;
MODEL TL = LINF*(1-EXP(-K*(AGE-t0)));
BY REPLICATE;
OUTPUT OUT = FITTED;
RUN;
I would like to create a dataset that contains parameter estimates for each replicate. The dataset would contain the following variables:
REPLICATE LINF K t0
and would look like this, for example:
1 568 .25 .02
2 520 .27 .14
3 491 .32 .01
etc.
I have also attached the SAS file.
Any help would be appreciated.
Try adding the statement:
ODS OUTPUT ParameterEstimates=samplePE;
to your NLIN procedure.
PG
PG,
That was helpful in summarizing the parameter estimates by replicate. However, is there a way to get them into a format that is easier to work with? Ideally this would be a SAS data set.
Actually I may be ok working with it by exporting to Excel.
In the example I gave, samplePE would be a SAS dataset containing the parameter estimates. That seemed to be what you wanted.
Got it. Thanks!
Are you ready for the spotlight? We're accepting content ideas for SAS Innovate 2025 to be held May 6-9 in Orlando, FL. The call is open until September 25. Read more here about why you should contribute and what is in it for you!
Learn the difference between classical and Bayesian statistical approaches and see a few PROC examples to perform Bayesian analysis in this video.
Find more tutorials on the SAS Users YouTube channel.