topic Re: Bootstrap sampling in SAS Data Management

Bootstrap sampling

Rasheed — Fri, 01 May 2015 15:14:57 GMT

I wrote a program for bootstrap sample generation

It works fine but It has one issue

when I run that SAS program twice then newer bootstrap samples are merged with previously generated bootstrap samples in SAS output

E.g, First I generated 100 bootstrap sample and calculated 95 percentile from it

But when I again run that program then SAS out put show 200 bootstrap samples therefore 95 percentile is misleading in this case

Please help me how to resolve this issue

Re: Bootstrap sampling

Steelers_In_DC — Fri, 01 May 2015 15:23:28 GMT

I doubt anyone will have any valuable input without being able to look at your code.

Re: Bootstrap sampling

ballardw — Fri, 01 May 2015 17:10:26 GMT

It sounds like somewhere you are appending data to a dataset. The first step would be either to deleted the existing dataset or direct your output to a new dataset.

Other than that, details are needed to provide more specific help.

If you have code look for some place with Proc Append, Proc Datasets with append, a data step with 2 or more datasets on a Set statement or Proc SQL with a Union join.

Re: Bootstrap sampling

Rasheed — Sun, 03 May 2015 05:58:54 GMT

http://www.math.uni-bremen.de/~pigeot/software/engl_boot_sas.html

I am using this program plz help me in this regard

Re: Bootstrap sampling

Rasheed — Sun, 03 May 2015 06:00:27 GMT

here I am inserting a link of program that I am using

I would be really grate full to you if you could mention me the correction in it

http://www.math.uni-bremen.de/~pigeot/software/engl_boot_sas.html

Re: Bootstrap sampling

Rasheed — Sun, 03 May 2015 06:04:31 GMT

http://www.math.uni-bremen.de/~pigeot/software/engl_boot_sas.html

Re: Bootstrap sampling

Reeza — Sun, 03 May 2015 20:18:05 GMT

Add a parameter to your program to ensure that it creates a uniquely named data set for each iteration/call.