# Inner product of random Gaussian vector

architect
Hi,

I would like to ask a question please.

Assume we have a random vector $$X$$ that is distributed under the Gaussian model and take the inner product of this vector and another constant vector $$d$$. Will the source distribution (Gaussian) remain the same? My intuition (although I might be wrong) says yes since the inner product may be regarded as a scaling operation. I know that this is true when we multiply a random variable with a constant, I just wonder if it applies to the inner product as well.

BR,

Alex.

Last edited:

## Answers and Replies

SW VandeCarr
Hi,

I would like to ask a question please.

Assume we have a random vector $$X$$ that is distributed under the Gaussian model and take the inner product of this vector and another constant vector $$d$$. Will the source distribution (Gaussian) remain the same? My intuition (although I might be wrong) says yes since the inner product may be regarded as a scaling operation. I know that this is true when we multiply a random variable with a constant, I just wonder if it applies to the inner product as well.

BR,

Alex.

By definition if a random vector x$$=x_1+x_2+....+x_n$$ has a multivariate normal distribution, then the inner product y=<a,x>, where a is a constant vector:

y$$=a_{1}x_{1}+a_{2}x_{2},+.....+a_{n}x_{n}$$ is normally distributed.

Last edited:
architect
Thanks for your help. Appreciated!

BR,

Alex