Mplus Discussion >> Bayesian Regression parameter constraint

Topics
Last Day
Last 3 Days
Last Week
Tree View

Edit Profile


Bayesian Regression parameter constraint

Mplus Discussion > Latent Variable Mixture Modeling >

Message/Author

Puneet Tiwari posted on Tuesday, March 22, 2016 - 10:17 am

Is it possible to provide positivity constraint (beta>0) to some of the regression coefficients in bayesian linear regression or bayesian mixture regression?

Tihomir Asparouhov posted on Tuesday, March 22, 2016 - 2:22 pm

Yes

model: Y on X (beta);
model constraint: beta>0;

Puneet Tiwari posted on Wednesday, March 23, 2016 - 2:01 am

Thanks Tihomir,
But if we have more than one covariates:
Does this mean:

model: y on x1
x2 (beta2)
x3 (beta3);
model constraint: beta2>0; beta3>0;

Tihomir Asparouhov posted on Wednesday, March 23, 2016 - 4:57 pm

Yes