3 ms·
> This stochastic process is a Gaussian process if for each t in R X_t has Gaussian distribution. This is not quite strong enough. You need (X_t1, X_t2, ..., X
by CrazyStat 7y ago
> This stochastic process is a Gaussian process if for each t in R X_t has Gaussian distribution.
This is not quite strong enough. You need (X_t1, X_t2, ..., X_tn) to jointly have a multivariate Gaussian distribution for every finite (t1, ..., tn).
- graycat 7y agoUsually! I gave a weak definition! Also, for correlation, need to assume the expectations exist.