Timeline for What is theta in a negative binomial regression fitted with R?
Current License: CC BY-SA 3.0
7 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jan 31, 2022 at 19:51 | comment | added | Aniko | The only restriction is Theta > 0, and the value of 1 does not correspond to any case of special interest. The negative binomial distribution is always overdispersed compared to the Poisson, with smaller values of theta corresponding to more overdispersion [in this specific parametrization] | |
| Dec 20, 2021 at 7:37 | comment | added | MLE | What is the range of theta? Does theta have to be a value greater than one? | |
| Jun 16, 2014 at 18:05 | history | edited | gung - Reinstate Monica | CC BY-SA 3.0 | de-italicized "skew" |
| May 12, 2011 at 8:35 | vote | accept | MarkDollar | ||
| May 6, 2011 at 21:28 | comment | added | MarkDollar | You just have to read the first page. So theta (or k in anscombe) is the shape parameter of the negbin distribution and it manages, if the distribution is closer to gamma (k -> 0) or poisson (k -> infinity). But what does it mean to the fit? How can I interpret theta for example for the cars estimation? | |
| May 6, 2011 at 21:20 | comment | added | MarkDollar | thanks so far! This is good help... But: How can I interpret high or (low) values of theta? In McCaullaughs book generalized linear models there is a link to this paper from anscombe to make an interpretation of k. But unfortunately I don't really get it. The paper is claremontmckenna.edu/facultysites/math/FacMember/MOneill/… | |
| May 6, 2011 at 20:28 | history | answered | Aniko | CC BY-SA 3.0 |