Re: Correlation coefficient: proving it's between -1 and 1



Benjamin wrote: 
How do you go about showing this?

rho = cov(X,Y) / (sigmaX)(sigmaY)


Presuming you mean |rho| <= 1.

Isn't this a direct conclusion from applying the
Cauchy-Schwartz-Bunyakovskii inequality to the
definition of rho? Google this:

	cauchy-schwarz "correlation coefficient"

to see lots of hints and maybe even a simple
proof.

Dale
.

Relevant Pages