Page 2 of Congdon's "Applied Bayesian Modeling" has a brief discussion of this issue. The main problem is that the flat prior is improper and may create identification problem. A normal prior with wide range (large variance) is minimally informative and proper.
1 comment:
Page 2 of Congdon's "Applied Bayesian Modeling" has a brief discussion of this issue. The main problem is that the flat prior is improper and may create identification problem. A normal prior with wide range (large variance) is minimally informative and proper.
Post a Comment