+ Reply to Thread
Results 1 to 5 of 5

Thread: Model Ratio using Linear Regression

  1. #1
    Points: 14, Level: 1
    Level completed: 27%, Points required for next Level: 36

    Posts
    2
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Model Ratio using Linear Regression




    Hello, I have a question about linear regression modeling on ratio.

    The original model with dependent variable of ratio is Y= (Y1/ Y2). Both Y1 and Y2 are dollar amount represent sales and balance respectively. The dependent variable Y is a continuous variable between 0 and 1.

    The equation is Y=(Y1/Y2)=a+B1X1+B2X2. Basically this is a simple linear regression model.

    My question is after I get the predicted Y by using the simple regression model, can I use the following equation to get the predicted Y1, the sales.

    Y1=Y*Y2, where Y2 is the future estimated balance.

    If this is correct, however, I see the predicted Y1 will eventually goes to 0, since Y2, the balance goes to 0 eventually. But Y1, the sales, can never goes to 0 in practical world. I am not sure where it goes wrong.

    Thank you for your help.

  2. #2
    Points: 4,664, Level: 43
    Level completed: 57%, Points required for next Level: 86
    kiton's Avatar
    Location
    Corn field
    Posts
    234
    Thanks
    47
    Thanked 51 Times in 46 Posts

    Re: Model Ratio using Linear Regression

    Note, linear regression requires the assumption of unbounded DV, whereas in your case the ratio is bounded between 0 and 1. As such, your estimates might be biased. To address this issue, you can explore beta regression or glm fractional logit. Now, answering your question, calculating Y1 does not seem plausible to me.

  3. #3
    Omega Contributor
    Points: 38,374, Level: 100
    Level completed: 0%, Points required for next Level: 0
    hlsmith's Avatar
    Location
    Not Ames, IA
    Posts
    6,998
    Thanks
    398
    Thanked 1,186 Times in 1,147 Posts

    Re: Model Ratio using Linear Regression

    kiton,


    Two comments:


    You live in a corn field, I love it. We will hope sweet corn and not feed corn.


    Second, I had not heard of fractional logit. Do you have any more information before I check the web?


    Thanks!


    P.S., I just found this:


    http://support.sas.com/resources/pap.../1304-2014.pdf
    Stop cowardice, ban guns!

  4. #4
    Omega Contributor
    Points: 38,374, Level: 100
    Level completed: 0%, Points required for next Level: 0
    hlsmith's Avatar
    Location
    Not Ames, IA
    Posts
    6,998
    Thanks
    398
    Thanked 1,186 Times in 1,147 Posts

    Re: Model Ratio using Linear Regression

    Still not sure these approaches get at your question. You could always say such in such ratio to solve for Y1, but certain values for your Y1 or Y2 may not be observed in your data, so they may not be identifiable or possible for extrapolation???
    Stop cowardice, ban guns!

  5. #5
    Points: 14, Level: 1
    Level completed: 27%, Points required for next Level: 36

    Posts
    2
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Re: Model Ratio using Linear Regression


    Thank you very much for all of you. I have solved the problem. It is because certain values for Y1 or Y2 not be observed in my data.

+ 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