I want to know how to write the $\chi^{2}$ distance between two multivariate Gaussian distributions $f$ and $g$ in terms of their parameters only. The parameters of $f$ is the vector $\mu_{1}$ and a covariance matrix $\Sigma_{1}$. The parameters of $g$ is the vector $\mu_{2}$ and a covariance matrix $\Sigma_{2}$.
Asked
Active
Viewed 173 times
1

gung - Reinstate Monica
- 132,789
- 81
- 357
- 650

alya
- 31
- 2
-
I would think that distance would be a measure of the separation of the mean vectors. But that would require a common scale. If the covariance matrices were equal I would think the Mahalanobis distance might be what the OP is referring to. – Michael R. Chernick Aug 04 '12 at 14:30
-
3That isn't the best choice of duplicate. This should be a duplicate of [What is the distribution of the euclidean distance between two normally distributed random variables](http://stats.stackexchange.com/q/9220/7290) instead. – gung - Reinstate Monica Dec 05 '15 at 15:50
-
@gung: agreed; let's close this 33666 and that 30218 as duplicates of the older 9220. – Felipe G. Nievinski Dec 05 '15 at 18:54