Help! How to give multiplicative error in turing or DifferentialEvolutionMCMC?

New to Julia
,
1

I used to work in r and now i’m switching to Julia. In r, I was using DifferentialEvolution package where its very straight forward if you want to give multiplicative or additive noise with respect to data as shown:

likelihood1 = sum(dnorm(x= prediction, mean= data, sd=(data+0.01)*SD,log=TRUE))

I tried same logic in DiffenrentialEvolutionMCMC in julia as it has a direct loglikelihood function but its not working.

for j in 1:3
for i in 1:6

   LL += logpdf(MvNormal(predicted[:,i], data[j][:,i] .*σ ), data[j][:,i])
  
  end
end

Also I’m trying in turing.jl

for j in 1:3
for i in 1:6

   data[j][:,i] ~ MvNormal(predicted[:,i], data[j][:,i] .*σ)
 

  end  
end

I’m pretty sure there is something basic i’m missing here… It will be really nice if someone can help… Thank You :slight_smile: