I have run the model, and it successfully compiled all 4 MCMC chains.
I need to see its plot, here is my code:
summary(fit_WTA_CFC1) # This, however, brought me more results than I expected - with #many ‘mu’. I used the family = categorical (link =“logit”). I don’t understand these, and I #don’t know whether the error comes from these.
plot(conditional_effects(fit_WTA_CFC1), ask = TRUE) #here is the error it brought:
Error in h(simpleError(msg, call)) :
error in evaluating the argument ‘x’ in selecting a method for function ‘plot’: Please set ‘categorical’ to TRUE.
That’s not quite enough information for us to help debug. Can you provide a reproducible example of the problem you’re having (by which I mean enough data and model to reproduce your output)?
My suggestion is to start by reading up on GLMs and what the family and link entails in terms of the actual model. The categorical family says it’s a multi-category categorization and the logit link says how to convert unconstrained linear predictors into probabilities. I’m not sure which parameterization of multi-logic is used by brms. The Stan users’ guide goes over a couple possible ways to code these models.