BookmarkSubscribeRSS Feed
roberto93
Calcite | Level 5

Hi! I am using PROC HPGENSELECT with weights but results are different than in other analysis tools. Could you tell which type of weight is used in this PROC (f.e. frequency weights or probability weights)? Is it possible to use importance weights in this PROC? Thanks for help.

5 REPLIES 5
Rick_SAS
SAS Super FREQ

The GENSELECT procedure (in fact, all regression procedures except SURVEY* procedures) uses analytical weights, also called n inverse variance weight. See the following articles:

 

The difference between frequencies and weights in regression analysis

Visualize a weighted regression

How to understand weight variables in statistical analyses

 

The last article discusses 'importance weights.' The GENSELECT procedure supports frequencies and analytical weights.

roberto93
Calcite | Level 5

Thank You for answer. However it's pretty sad that SAS doesn't support one the most common types of weights. Is there a way to use at least the probability weight in this PROC?

Rick_SAS
SAS Super FREQ

If you specify your model and post some sample data, someone will be able to advise you as to how to fit the model in SAS.

StatDave
SAS Super FREQ

If you have survey data with weights, then that is supported by our set of SURVEY procedures such as PROC SURVEYREG, PROC SURVEYLOGISTIC, and others. See those procedures to see if your needs are covered by one of those procedures. 

roberto93
Calcite | Level 5

Well, I want to build Zero Inflated models, (based on stepwise). Using weights, recived parameters would be better. I don't think that PROC SURVEYREG will be enough for that.

SAS Innovate 2025: Save the Date

 SAS Innovate 2025 is scheduled for May 6-9 in Orlando, FL. Sign up to be first to learn about the agenda and registration!

Save the date!

What is ANOVA?

ANOVA, or Analysis Of Variance, is used to compare the averages or means of two or more populations to better understand how they differ. Watch this tutorial for more.

Find more tutorials on the SAS Users YouTube channel.

Discussion stats
  • 5 replies
  • 1982 views
  • 0 likes
  • 3 in conversation