Say we have two multivariate normal random variables $X$ and $Y$ with the same dimensionality, means $\mu_x$ and $\mu_y$, and covariance matrices $\Sigma_x$ and $\Sigma_y$. It is possible to show that the distribution of $X+Y$ is also multivariate random normal with mean $\mu_x + \mu_y$ and covariance $\Sigma_x+\Sigma_y$, but what if the distributions are not equally weighted?

Formally, what is the distribution of the sum of $p_xX+(1-p_x)Y$ where $p_x\in(0,1)$ is the proportion of the sum attributable to the distribution $X$?

  • $\begingroup$ What's the distribution of $p_xX$? $\endgroup$
    – Bananach
    Aug 4, 2021 at 23:13
  • $\begingroup$ Are $X$ and $Y$ independent of each other? $\endgroup$
    – Henry
    Aug 5, 2021 at 0:56

1 Answer 1


Any linear combination of independent multivariate normals of the same dimension is also multivariate normal. Therefore, if we assume $X$ and $Y$ are independent then

$$Z_p = pX + (1-p)Y \sim \mathcal{N}\left(p\mu_X + (1-p)\mu_Y, p^2\Sigma_X+ (1-p)^2\Sigma_Y\right)$$


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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