Skip to content

Covariance and Correlation Coefficient ​

Markov-inequality ​

Let g:R→R be a monotonically increasing, positive function, X≥0 a random variable, for which E(X)<∞ and ϵ>0. Then

P(X≥ϵ)≤E(g(X))g(ϵ).

Especially, if g(x)=x, then

P(X≥ϵ)≤E(X)ϵ.

Chebyshev-inequality ​

Let X be an arbitrary random variable, for which Var(X)<∞ and ϵ>0. Then

P(|X−E(X)|≥ϵ)≤Var(X)ϵ2.

Covariance ​

Covariance of two random variables:

cov(X,Y)=E(XY)−E(X)E(Y).

The joint density function of (X,Y) is f(x,y) if for every appropriate subset A⊆R2 we have

P(X∈A)=∫Af(x,y)dxdy.

Calculation of covariance if the joint density function is f(x,y):

cov(X,Y)=∫∗−∞∞∫∗−∞∞xy⋅f(x,y)dxdy−(∫∗−∞∞x⋅fX(x)dx)⋅(∫∗−∞∞y⋅fY(y)dy),

where fX(x) and fY(y) are the marginal densities that can be computed as fX(x)=∫−∞∞f(x,y)dy and fY(y)=∫−∞∞f(x,y)dx.

If random variables X and Y are independent, then cov(X,Y)=0. On the other hand, it may happen that cov(X,Y)=0, but X and Y are not independent.

Correlation coefficient ​

The (Pearson) correlation coefficient of two random variables X and Y

R(X,Y)=cov(X,Y)sd(X)sd(Y).

(this is always at least −1 and at most 1)

R(X,Y)=1 means that Y=aX+b for some a,b, with a>0.