Programming the statistical procedures from SAS

Best technique to choose for model comparison

Accepted Solution Solved
Reply
Contributor
Posts: 24
Accepted Solution

Best technique to choose for model comparison

Hello! Which statistical technique is used for model selection from the regression model (multiple linear regression). Adjusted R Square or Mallow's CP or Average Squared Error? which one is the most reliable one while choosing the model? Sameer Shaikh

Accepted Solutions
Solution
‎11-30-2016 05:17 AM
Grand Advisor
Posts: 9,444

Re: Best technique to choose for model comparison

There are many statistic you can use. AIC,BIC , ASE, .........
Check documentation of proc glmselect . the most popular method is LOESS .



View solution in original post


All Replies
Grand Advisor
Posts: 16,850

Re: Best technique to choose for model comparison

AIC. 

 

There isn't a 'best technique' and the options you've listed are a small subset of the options. 

Grand Advisor
Posts: 9,444

Re: Best technique to choose for model comparison

The modern method is using Training data,Validated data, Testing data.
Check PROC GLMSELECT .

Contributor
Posts: 24

Re: Best technique to choose for model comparison

Ksharp and Reeza,

 

Thanks for the reply. On validation data which method is useful? Radjusted or mallowco or RMSE?

 

 

Solution
‎11-30-2016 05:17 AM
Grand Advisor
Posts: 9,444

Re: Best technique to choose for model comparison

There are many statistic you can use. AIC,BIC , ASE, .........
Check documentation of proc glmselect . the most popular method is LOESS .



☑ This topic is SOLVED.

Need further help from the community? Please ask a new question.

Discussion stats
  • 4 replies
  • 112 views
  • 0 likes
  • 3 in conversation