- Operating System: Windows10
- brms Version:2.8
I created a nonlinear mixed-effects model.
Now I would like to generate a plot with my actual data points and a curve with parameter values, which are the parameter estimates obtained when fitting the model.
However, if I only plot these, I lose a lot of information about the uncertainty of the parameters.
So I was wondering what the best way is to represent this uncertainty?
A) Should I just draw samples from the distributions of the parameters (they all have normal prior, so maybe using multivariate normal?) and, generate, say, 100 curves with these, and plot them at the same plot as my fitted curve?
B) Or is there any way to plot error bands to represent that uncertainty using brms?
Or should I just use the estimate I got for Sigma (in my case family=gaussian()) ´,
even though in this case I would lose information about the uncertainty of Sigma?
I would really appreciate it if someone could give me some insight.