We have a multivariable (2 variables) parameter which has a (asymptotically) multivariate normal distribution, with known mean and covariance matrix.
How do we find the confidence interval for just one of the two variables?
Do we need to incorporate covariance?
e.g (a,b) ~ N(0,\Sigma)
Is the confidence interval for a:
[math]a = 0 \pm 1.96* \Sigma_{1,1}[/math]
Confidence intervals don't make sense in 2D, use confidence ellipses instead.
Also >>8969578