+ Reply to Thread
Results 1 to 1 of 1

Thread: Automatically finding relationships between 3 or more variables

  1. #1
    Points: 13, Level: 1
    Level completed: 25%, Points required for next Level: 37

    Posts
    2
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Automatically finding relationships between 3 or more variables




    Hello,

    Total newbie here - but I'm hoping someone can help me out.
    I have finished running a multiple regression analysis and have my results. However, I want to go back and see if there any unusual relationships between variables and my results that I might have missed.

    For example:
    Case# - Area - Type - Ratio of Predicted to Actual.
    1 - North - A - 0.5
    2 - North - A - 0.5
    3 - North - A - 0.5
    4 - North - B - 1.75
    5 - North - B - 1.75
    6 - South - B - 0.5
    7 - South - B - 0.5
    8 - South - B - 0.5
    9 - South - A - 1.75
    10 - South - A - 1.75

    I'm trying to get the ratio as close to 1 as possible for each case/group. The average ratio for each variable is 1 (The average for either North/South is 1, the average for each type is also 1). However, the average for each Type within a certain Area is not 1 (e.g. the average of A within the North is 0.5). I need a way to find these unusual combinations so I can identify flaws in my model.

    In reality, I have a database with over 50 variables(many of which are continuous), and 5000+ cases. Blindly testing different combinations of variables would be extremely time-consuming.

    Is there some way in SPSS, or in another program, to automatically find the combination of variables where the average ratio is far away from 1? Preferably, some way to automatically test every possible combination of variables.

    Thanks
    Last edited by Tzo; 10-04-2015 at 03:38 PM.

+ Reply to Thread

           




Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts






Advertise on Talk Stats