Timeline for What is theta in a negative binomial regression fitted with R?
Current License: CC BY-SA 3.0
6 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jan 20, 2020 at 19:43 | comment | added | BurninLeo | If you have no idea what theta is, try this fit = glm.nb(y ~ 1, data=data); fit$theta; | |
| Jun 16, 2014 at 18:40 | comment | added | Momo | Thanks for this clarification. I guess one problem for this misunderstandings comes from the R output of the sort where it reads "(Dispersion parameter for Negative Binomial(theta) family taken to be 1)". Here dispersion refers to the $\phi$ in $cov(\hat{\beta})=\phi(X^T\hat{Ŵ}X)^{-1}$, which you seem to call scale. So many people opt for calling the $\theta$ differently. Also derived from the R function, I tended to call $\mu$ location and $\theta$ "shape"-the latter of which I don't find unreasonable as it certainly influences the shape. | |
| Jun 16, 2014 at 18:02 | history | edited | gung - Reinstate Monica | CC BY-SA 3.0 | added links & paragraph breaks; light editing & formatting |
| Jun 16, 2014 at 17:32 | history | edited | Scortchi♦ | CC BY-SA 3.0 | improved formatting |
| Jun 16, 2014 at 17:31 | review | First posts | |||
| Jun 16, 2014 at 17:38 | |||||
| Jun 16, 2014 at 17:13 | history | answered | Joseph Hilbe | CC BY-SA 3.0 |