This is the logistic distribution of single random variable (taken from Wikipedia).

x = random variable mu = mean of all random variables s = variance.

enter image description here

Now, I want to do a Bivariate logistic distribution (having two random variables x1 and x2). My dataset is going to be image pixel values!

When I find the covariance of two random variables, it turns out to be a 2 x 2 square matrix but I need a single number! How to actually compute a bivariate model?


1 Answer 1


The covariance of any multivariate distribution is always a matrix. Maybe it is of help to you to know what this matrix contains:

-In general, each entry i,j of the matrix gives you the correlation between the RVs i and j.

-So a diagonal element i,i will simply be the variance of the RV i.

-If RV i and j are independent, then the non-diagonal entry i,j is zero (they are uncorrelated). The reverse of this statement is not generally true, but it is for Gaussians iirc.

I don't know what you are trying to do, but perhaps the trace of the matrix or its determinant is of use to you?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.