Timeline for Bayesian Hierarchical Model - Exact Conjugate Solution?
Current License: CC BY-SA 3.0
8 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jun 2, 2014 at 17:44 | answer | added | user44764 | timeline score: 0 | |
| Jun 2, 2014 at 17:44 | answer | added | jaradniemi | timeline score: 1 | |
| Jun 2, 2014 at 1:33 | comment | added | Glen_b | That clarifying information should probably go in your question | |
| Jun 2, 2014 at 0:25 | comment | added | user1375871 | $\tau$ is the between studies precision which is unknown. $\tau_i$ is the within-study precision which is known from the data. | |
| Jun 1, 2014 at 20:18 | comment | added | Sean Easter | Do you have any other prior beliefs for $\tau_i$, or any assumptions about how it should be modeled? I can't quite suss out why there are two sources of variance. | |
| Jun 1, 2014 at 19:18 | comment | added | user1375871 | Yes $\tau_i$ is observed in each study. It is not assumed to be drawn from the same gamma | |
| Jun 1, 2014 at 18:40 | comment | added | Sean Easter | By the mixture property of the normal distribution, $Y_i | \theta_i$ is given by $N(\mu, \frac{1}{\tau} + \frac{1}{\tau_i})$. Is $\tau_i$ observed in each study, and assumed drawn from the same gamma distribution? | |
| Jun 1, 2014 at 18:17 | history | asked | user1375871 | CC BY-SA 3.0 |