I have been studying the error structure of Bayesian isotope source mixing models recently. When dealing with data containing only one mixture, simmr use a solo run. What is the error structure used when running a solo run? Process error + residual error or process error only (like MixSIR and MixSIAR)?
Hi, when running a solo run the model uses a prior for $\sigma_j$ with high prior mass on zero - so essentially it becomes process error only like MixSIR/MixSIAR
Dear Parnell,
I have been studying the error structure of Bayesian isotope source mixing models recently. When dealing with data containing only one mixture, simmr use a solo run. What is the error structure used when running a solo run? Process error + residual error or process error only (like MixSIR and MixSIAR)?
Sincerely, Liu