Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Linear Correlation

  1. Oct 15, 2007 #1
    I am going through some of the problems in a statistical physics book. I am stuck on the following question:

    If we have two random variables X and Y, related by Y=mX + b => y(i) = mx(i) + b, where b and m are deterministic, prove that corr(X,Y) = m/sqrt(m^2) = sgn(m), where sgn(m) is the sign function of m.

    I can see that this is perfect linear correlation, but I am not sure as to what is the obvious proof....I know that corr(X,Y) = (E(XY) - E(X)E(Y))/sd(X)*sd(Y)....

    Is the following derivation even logically valid?
    = E(X(mX+b)) - E(X)E(mX+b)
    =E(mX^2 + bX)-E(X)E(mX+b)
    =m(Var(X))+bE(X) +b

  2. jcsd
  3. Oct 15, 2007 #2


    User Avatar
    Science Advisor

    Your approach is valid. Your arithmetic is wrong.

    E(XY)-E(X)E(Y)=m(Var(X)). The b terms cancel when you do it right!

    The denominator will be |m|Var(X).
  4. Oct 16, 2007 #3
    Thank you!
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook