2

according to the definition

Cov(X,Y)=E[XY]-E(X)E(Y)

I happen to get a negative value, I guess there is a problem?

When I tried to get the correlation since the E(X) and E(Y) are really big I ended up with another negative value which cannot be correct because I have to squareroot the variance to get the sigma_x and sigma_y.

Would be glad if someone could tell me what could have went wrong....

natsu
  • 354

1 Answers1

3

Covariance (and correlation coefficient) can be negative.

If the covariance between $X$ and $Y$ is positive then $Y$ tends to increase as $X$ increases, and if the covariance is negative then $Y$ tends to decrease as $X$ increases.

M.B.M.
  • 5,406