Return the posterior on beta given a prior (g) that is
a mixture of normals (class normalmix) and observation
$betahat ~ N(beta,sebetahat)$
Usage
posterior_dist(g, betahat, sebetahat)
Arguments
g
a normalmix with components indicating the prior; works
only if g has means 0
betahat
(n vector of observations)
sebetahat
(n vector of standard errors/deviations of
observations)
Value
A list, (pi1,mu1,sigma1) whose components are each k by n matrices
where k is number of mixture components in g, n is number of observations in betahat