I have a pretty basic mixture regression thus:
. fmm 2: regress y x1 x2, lcprob(x1)
What I would like to do is constrain the coeficient x1 across classes. I look at by e(b) matrix and find:
y: y:
1.Class# 2.Class#
c.x1 c.x1
2.0473413 1.897758
I therefore create a constraint:
. constraint 1 1.Class#c.x1 = 2.Class#c.x1
and then fit the fmm like this:
. fmm 2, constraints(1): regress y x1 x2, lcprob(x1)
But that parameter does isn't constrained and I get no error. In fact, I can't get any constraints on x1 to work. Perhaps I'm using the incorrect label. Is my constraint syntax correct?
Thanks so much
. fmm 2: regress y x1 x2, lcprob(x1)
What I would like to do is constrain the coeficient x1 across classes. I look at by e(b) matrix and find:
y: y:
1.Class# 2.Class#
c.x1 c.x1
2.0473413 1.897758
I therefore create a constraint:
. constraint 1 1.Class#c.x1 = 2.Class#c.x1
and then fit the fmm like this:
. fmm 2, constraints(1): regress y x1 x2, lcprob(x1)
But that parameter does isn't constrained and I get no error. In fact, I can't get any constraints on x1 to work. Perhaps I'm using the incorrect label. Is my constraint syntax correct?
Thanks so much