At first I have a data frame that contains 1020 devices, named by Serial_number, and each device consists of several measurements (Amplification against Voltage).
I want to apply a Bayesian mixed model but don't know how to choose the priors.
library(blme) #bayesian linear mixed effects package
summary(blme34.fit <- blmer(log(log(Amplification)) ~ poly(Voltage,3) + (poly(Voltage,3) |Serial_number), APD,
cov.prior = gamma))
This works in principle but I get similar convergence warnings as I did for lmer-models (lme warning message because of random effects). Therefore I assume I need some different priors but all examples from the official package description (https://cran.r-project.org/web/packages/blme/blme.pdf) do not work in a way that errors are produced ("formal argument "cov.prior" matched by multiple actual arguments*", "*'nrow' or 'ncol' cannot be specified when 'x' is a matrix").
How can I add priors?
