Closed ga72kud closed 2 years ago
I am wondering if there are also possibilities to use multivariate distributions. Something like that:
μ=[1.;.1] Σ=[1. 0.; 0. 1.] N=MvNormal(μ, Σ) μₐ=[2.;.4] Σₐ=[1. 0.; 0. 1.] Nₐ=MvNormal(μ, Σ) Omega.replace(N, μ=>Nₐ)
My second questions is Omega.jl only valid for singlevariate normal distributions:
a=Distributions.Categorical([.8, .2]) rand(a) N=Normal(μ₁, σ₁) Omega.replace(N, μ₁=>a)
I could find something like that. and use the replacement command for a and b. It could also work for Sigma, if each value in the covariance matrix is sampled by a Normal distribution
a=Normal(.1, .1) b=Normal(.2, .1) μ=[mean(a);mean(b)] Σ=[1. 0.; 0. 1.] N=MvNormal(μ, Σ)
I am wondering if there are also possibilities to use multivariate distributions. Something like that:
My second questions is Omega.jl only valid for singlevariate normal distributions:
I could find something like that. and use the replacement command for a and b. It could also work for Sigma, if each value in the covariance matrix is sampled by a Normal distribution