Timeline for Should one use the same overdispersion parameter when comparing Binomial models?
Current License: CC BY-SA 3.0
6 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Aug 20, 2016 at 1:45 | vote | accept | Nik Tuzov | ||
| Aug 19, 2016 at 14:37 | vote | accept | Nik Tuzov | ||
| Aug 19, 2016 at 14:38 | |||||
| Aug 18, 2016 at 2:17 | comment | added | Jonny Lomond | The negative binomial distribution allows you to fit the dispersion parameter directly through maximum likelihood estimation, and contains the Poisson distribution as a special case (when the dispersion is 1). The key point is that all parameters are fit so as to maximize the likelihood of the model. | |
| Aug 17, 2016 at 21:58 | comment | added | Nik Tuzov | What about Negative Binomial model? It appears that, once estimation of the dispersion parameter is built into the procedure, everyone re-estimates dispersion automatically w/o thinking about it. | |
| Aug 17, 2016 at 11:15 | history | edited | Jonny Lomond | CC BY-SA 3.0 |
deleted 3 characters in body
|
| Aug 16, 2016 at 16:02 | history | answered | Jonny Lomond | CC BY-SA 3.0 |