How to group by in a JSL fit model script?

[mattdh88]

linfit=Fit Model(

      Y( :PT ),

      Effects( :Pts 2 ),

      Personality( Standard Least Squares ),

      Run model

);

linfit<<prediction formula;

I have a simple script using a fit model that saves the prediction formula to my current data table. It looks at a basketball player's points scored on a given night vs. what the opponent's defense typically allows. In the table, there are multiple players and duplicate values for each player. Instead of fitting a linear regression for all players and using one prediction formula, I'd like to fit a linear regression for each individual player, so that there are multiple prediction formulas within one column. When I let JMP create the script for me, the script just runs over and over with a "where player = X" clause for each player. The players change on a regular basis so this can't be used. I thought I could just add a "Group by(:players)," line after Effects(:Pts 2) but that isn't grouping anything for me. How do I go about using group by in a fit model script?

[PGStats]

Please post your question to the JMP community forum

https://community.jmp.com/welcome

PG