I need the following distribution as a prior for a model:
hist( rlnorm(1000,0,0.5)-5 )
Currently, I solve this problem by having the -5 subtraction in the model formula:
formula ← bf(y ~ (plateau-5) + t0 * exp(-time / tau),
plateau + t0 + tau ~ 1,
nl = TRUE)
prior ← prior(lognormal(0, 0.5), nlpar = “plateau”)
However, this complicates later modifications of the prior (i.e., in the setting of an adaptive fitting function taking other variables into account), and I always have to note that the actual coefficient-of-interest is (estimated-coefficient - 5).
Is there a way of moving the ‘-5’ away from the formula into the prior?
Thank you for your help in advance.