cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 
  • JMP 19 is here! See the new features at jmp.com/new.
  • Due to global connectivity issues impacting AWS Services, users may experience unexpected errors while attempting to authorize JMP. Please try again later or contact support@jmp.com to be notified once all issues are resolved.

Discussions

Solve problems, and share tips and tricks with other JMP users.
Choose Language Hide Translation Bar
viskovicz00012
Level II

Outlier Deletion

Is there a way to script the deletion of extreme outliers per column? My objective is once i run the script, it concatenates data tables, delete extreme outliers per column and display graphs but im having some sort of problem with the outlier part.
1 ACCEPTED SOLUTION

Accepted Solutions
tom_abramov
Level V

Re: Outlier Deletion

I use Jackknife Distances for outliers.

For extreme outliers I use upper limit = 6.

In the following example I use upper limit = 2.

 

 

Names Default To Here( 1 );
dt = Open( "$SAMPLE_DATA/Big Class.jmp" );
JackknifeObj = dt << Multivariate( Y( 4 ), Jackknife Distances( 1, Save Jackknife Distances( 1 ) ) );
JackknifeObj << Close Window();
dt << Select Where( :Jackknife Distances > 2 );
dt << Hide and Exclude << Clear Select << Delete Column( Jackknife Distances );
 
 
 

View solution in original post

2 REPLIES 2
tom_abramov
Level V

Re: Outlier Deletion

I use Jackknife Distances for outliers.

For extreme outliers I use upper limit = 6.

In the following example I use upper limit = 2.

 

 

Names Default To Here( 1 );
dt = Open( "$SAMPLE_DATA/Big Class.jmp" );
JackknifeObj = dt << Multivariate( Y( 4 ), Jackknife Distances( 1, Save Jackknife Distances( 1 ) ) );
JackknifeObj << Close Window();
dt << Select Where( :Jackknife Distances > 2 );
dt << Hide and Exclude << Clear Select << Delete Column( Jackknife Distances );
 
 
 
txnelson
Super User

Re: Outlier Deletion

@tom_abramov approach is a good method for identifying outliers. Are you looking for what method to use to identify outliers or do you already have a method of determining the outliers that you want to use?

 

 

Jim

Recommended Articles