The new forums will be named Coin Return (based on the most recent vote)! You can check on the status and timeline of the transition to the new forums here.
The Guiding Principles and New Rules document is now in effect.

Least Squares Fit and Residuals Questions

StudioAudienceStudioAudience Registered User regular
edited March 2007 in Help / Advice Forum
I'm working on a problem set right now, and I'm sort of stump on the following queries. I'm not looking for answers, but tips/hints would be awesome.


Given x1,x,2,x3,x4, and y1,y2,y3,y4 as data points, let yhat = a1(xi) + a0, i = 1,2,3,4 be the estimates of the least squares fit.

The problem wants me to show that the mean of the residuals (yhat - y) is equal to 0.

I have:

[(a1x1 + a0 - y1) + (a1x2 + a0 - y2) + (a1x3 + a0 - y3) + (a1x4 + a0 - y4)]/4 as being equal to the mean. I guess you can simplify this to:

[a1(x1+x2+x3+x4) + 4a0 - (y1+y2+y3+y4)]/4

However, this is as far as I can get before I'm stumped. How would I go about showing that this is equal to 0 when no definite values are given for any of the variables?

studaud.png
GFWL: studaud (for SF4)
StudioAudience on

Posts

  • RocketScienceRocketScience Registered User regular
    edited March 2007
    You haven't taken into account how a Least Squares Fit works, ie by calculating the constants a0 and a1 such that the sum of the residuals is minimised.

    Hint
    http://mathworld.wolfram.com/LeastSquaresFitting.html
    The answer to your problem is hidden somewhere between equation (1) and (6).

    RocketScience on
  • StudioAudienceStudioAudience Registered User regular
    edited March 2007
    Ah, gotcha. Thanks, that makes sense now.

    StudioAudience on
    studaud.png
    GFWL: studaud (for SF4)
Sign In or Register to comment.