+ Reply to Thread
Results 1 to 7 of 7

Thread: Intercept becoming negative(More IDVs)

  1. #1
    Points: 277, Level: 5
    Level completed: 54%, Points required for next Level: 23

    Posts
    8
    Thanks
    4
    Thanked 0 Times in 0 Posts

    Intercept becoming negative(More IDVs)




    Hi All,
    I have a small query.

    When regressing revenue against cost, my intercept is positive for X% combos where each combo consists of data points at weekly level for Region x Product x Channel. When i am adding more variable into the regression equations, i see many combos now having negative intercept when regressing revenue against cost and a few more IDVs.

    Can someone explain as to why this is happening and what can be inferred from it??

    Thanks & Regards

  2. #2
    Omega Contributor
    Points: 38,396, Level: 100
    Level completed: 0%, Points required for next Level: 0
    hlsmith's Avatar
    Location
    Not Ames, IA
    Posts
    7,001
    Thanks
    398
    Thanked 1,186 Times in 1,147 Posts

    Re: Intercept becoming negative(More IDVs)

    Do you know what the intercept represents?

    Can your dependent variable take on a negative value?

    Does your dependent variablr take on a negative value in the dataset?
    Stop cowardice, ban guns!

  3. #3
    Points: 277, Level: 5
    Level completed: 54%, Points required for next Level: 23

    Posts
    8
    Thanks
    4
    Thanked 0 Times in 0 Posts

    Re: Intercept becoming negative(More IDVs)

    Well my DV is the revenue. It should not be negative. No we had no cases in the data set where in the DV i.e. the revenue was negative. Let's say i am running a regression with Revenue as DV and Cost as IDV then it would mean a negative renvenue in case the cost is zero.

  4. #4
    Omega Contributor
    Points: 38,396, Level: 100
    Level completed: 0%, Points required for next Level: 0
    hlsmith's Avatar
    Location
    Not Ames, IA
    Posts
    7,001
    Thanks
    398
    Thanked 1,186 Times in 1,147 Posts

    Re: Intercept becoming negative(More IDVs)

    A negative intercept means predicted y takes on a negative value when predictors equal 0 or reference value. Look up extapolation and check back here.
    Stop cowardice, ban guns!

  5. #5
    TS Contributor
    Points: 7,284, Level: 56
    Level completed: 67%, Points required for next Level: 66

    Location
    MD, USA
    Posts
    465
    Thanks
    8
    Thanked 11 Times in 11 Posts

    Re: Intercept becoming negative(More IDVs)

    Quote Originally Posted by vikash_124124 View Post
    my DV is the revenue. It should not be negative. we had no cases in the data set where in the DV i.e. the revenue was negative.
    Since this has been brought up, I've always wondered:
    Let's say you have a DV value that can't go negative but, e.g., you have only a few values so the unreliable regression line shows that it does.
    Both X & Y axes start at zero.

    Can you justify somehow causing a positive value for the DV such that it causes the regression line to go exactly through zero?
    It's what
    you know to be true vs. what your graph shows.

    If so, is this "fudging", or "sanitizing", or just plain illegal?

    TIA

  6. #6
    Omega Contributor
    Points: 38,396, Level: 100
    Level completed: 0%, Points required for next Level: 0
    hlsmith's Avatar
    Location
    Not Ames, IA
    Posts
    7,001
    Thanks
    398
    Thanked 1,186 Times in 1,147 Posts

    Re: Intercept becoming negative(More IDVs)

    I know you can do it. You can also run models without intercept. Though I am not versed in these things.

    Typically you cant generalize to data not in your variable 's range.
    Stop cowardice, ban guns!

  7. #7
    TS Contributor
    Points: 7,284, Level: 56
    Level completed: 67%, Points required for next Level: 66

    Location
    MD, USA
    Posts
    465
    Thanks
    8
    Thanked 11 Times in 11 Posts

    Re: Intercept becoming negative(More IDVs)


    If I correctly understand the problem:

    For X = 1 & 3, and Y = 1 & 7, m (the slope) = 3, and b (the Y intercept) = -2.
    Now add an X value = 0 and a Y value such that the intercept = 0.

    I haven't yet done the closed form solution, if it is possible, but by tweaking a spreadsheet for my simple example an added Y value of 0.8 gives an intercept of 0.000.

    I used
    http://hotmath.com/hotmath_help/topi...-best-fit.html

+ 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