Covariance and correlation coefficient

Click For Summary

Discussion Overview

The discussion revolves around the mathematical properties of covariance and the correlation coefficient, specifically focusing on proving the maximum value of the covariance and the bounds of the correlation coefficient. The scope includes theoretical aspects and potential proofs related to these statistical concepts.

Discussion Character

  • Technical explanation
  • Debate/contested

Main Points Raised

  • One participant questions how to prove that the maximum value of 2*cov(x,y) can equal var(x) + var(y).
  • Another participant emphasizes the importance of proving that the correlation coefficient, defined as cov(x,y)/(sigma(x)*sigma(y)), can only range between -1 and 1.
  • Some participants suggest that empirical testing with random and real data sets shows that the correlation coefficient's bounds are never violated.
  • Several participants inquire about the existence of a formal proof for the correlation coefficient's bounds.
  • One participant recalls having seen a proof related to this topic during their teaching of statistics but suggests that it can be found online.
  • Another participant expresses frustration with online resources and seeks visual aids to better understand the concepts, mentioning potential confusion related to dot products.
  • A later reply identifies the relationship between the correlation coefficient and the Cauchy-Schwarz inequality, stating that it is a special case of this inequality.

Areas of Agreement / Disagreement

Participants generally agree on the need for proofs regarding the correlation coefficient's bounds, but there is no consensus on the existence of a satisfactory proof or the best way to understand the concepts involved.

Contextual Notes

Some participants express uncertainty about the relationship between empirical observations and theoretical rules, indicating a potential gap in understanding or assumptions about the data sets used.

Josh S Thompson
Messages
111
Reaction score
4
How do you prove that the maximum value of 2*cov(x,y) can be is equal to var(x) + var(y).

Moreover, how do you prove that the correlation coefficient, cov(x,y)/(sigma(x)*sigma(y), can only be between -1 and 1.
 
Physics news on Phys.org
Forget the first sentence the question is only the second sentence
 
Try computing a bunch for random and real data sets and you will see that the rule is never violated.

Of course, I'm an experimentalist.
 
is there a proof?
 
Josh S Thompson said:
is there a proof?

I bet there is for a statement for which no counter example exists. I think I recall even seeing one when I taught statistics.

But you can probably Google it up as easily as I can.
 
Google sucks, I want some pictures bro. Because I did some examples and I don't understand, I think it doesn't violate those rules because of like dot products or something but I don't see the correlation coefficient. Can someone please enlighten me with some insight.
 

Similar threads

  • · Replies 43 ·
2
Replies
43
Views
6K
Replies
4
Views
1K
  • · Replies 4 ·
Replies
4
Views
2K
  • · Replies 1 ·
Replies
1
Views
3K
  • · Replies 17 ·
Replies
17
Views
3K
  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 13 ·
Replies
13
Views
5K
Replies
13
Views
2K
  • · Replies 42 ·
2
Replies
42
Views
6K
  • · Replies 1 ·
Replies
1
Views
2K