Specifying my mixture model

Interfaces brms
1

Hello I’m totally new on this forum.
I’m working recently with brms package and I try to modelize my data with this mixture gaussian model.
I’d like to fix intercept and linear coefficient equal for mu1 and mu2 .

I try to read documentation but I didn’t really find something about that or I didn’t understand.
Here’s my code

help
help774×312 12.5 KB

here’s my get prior

help3
help3589×527 6.63 KB

My problem is quite simple :I would like to have the same b and intercept prior for mu1 and mu2 because it’s the same value in my data.

If it’s possible I guess I have to change my formula but I don’t know how.

2

Welcome to the community! And, I’m terribly sorry nobody has replied in five days… :(

I would like to have the same b and intercept prior for mu1 and mu2 because it’s the same value in my data.

Haven’t you set the same prior on the Intercept mu1 and mu2, above, i.e., student-t(3,6.2,2.5) for both mu1 and mu2’s intercept?

If you do p <- get_prior(read4) you can then set the priors however you want by doing, e.g.,

p$prior[1] <- "lkj(2)"
1 Like
3

Thanks a lot for your answer !
I’m sorry I think I didn’t exposed my problem correctly.
I want to estimate b mu1 and b mu2 as the same parameter and then to have only one prior for both.
Same thing for intercept.
The problem to fix is then in the formula I think.
I want a mixture model not between 2 gaussian law but between the variance distribution of the same gaussian.

4

I’m sorry I don’t follow you. However, I’m sure @paul.buerkner might know :)