SPSS Statistics

 View Only
  • 1.  Setting a constant in TURF Analysis

    Posted Mon August 30, 2021 10:52 AM
    Is there a way to set a constant in TURF Analysis (e.g. I have 4 products and 1 of them must be in the assortment) using Statistics OOTB?  Is the best option to use weights, and assign a higher weight for the variable that I want to remain constant?  Or do you have another recommendation?
    Thanks!

    ------------------------------
    Liz Herrera
    ------------------------------

    #SPSSStatistics


  • 2.  RE: Setting a constant in TURF Analysis

    IBM Champion
    Posted Mon August 30, 2021 03:34 PM
    TURF provides a way to force certain variables to appear in every bundle.  From the TURF syntax help.

    FORCE specifies the variables that must be included in every combination. The value is a number, n, indicating that the first n variables listed in VARIABLES are required. The default value is 0 indicating that no specific variables are required. If FORCE=2 in the example above, q1 and q2 would always be included regardless of their effect on the reach. FORCE cannot be combined with removal cycles.

    --





  • 3.  RE: Setting a constant in TURF Analysis

    Posted Tue August 31, 2021 04:43 PM
    Hi Jon,
    Thanks for the response!  Would you recommend using the Python Essentials to do the TURF analysis then?  Or do you think it is something that could be updated in the extension that was embedded in Descriptive Stats (via the GUI) - in a future release?
    Thank you again,
    Liz

    ------------------------------
    Liz Herrera
    ------------------------------



  • 4.  RE: Setting a constant in TURF Analysis

    IBM Champion
    Posted Tue August 31, 2021 05:19 PM
    The SPSSINC TURF extension command is what the dialog box generates.  It perhaps isn't as clear in the dialog box as it might be and as it is in the syntax help, but the Number of Analyze Variables Always Required field is where you set this up if I understand what you want.

    From the dialog box help:
    Number of Analyze Variables Always Required The value, n, indicates the the first n variables listed in the Variables to Analyze box are required and will thus appear in every combination of variables regardless of their effect on the reach. If the value is 0, no variables are forced. If the value is positive, there cannot be any removal cycles.

    --





  • 5.  RE: Setting a constant in TURF Analysis

    Posted Tue August 31, 2021 08:30 PM
    I see now - you're right, the syntax help is clearer than the dialog box.  This is what I was looking for so thank you again Jon, appreciate the help!

    ------------------------------
    Liz Herrera
    ------------------------------