Timeline for interpreting output for glmmTMB for zero-inflated count data
Current License: CC BY-SA 4.0
6 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| May 18, 2020 at 16:15 | vote | accept | alessothegreat | ||
| May 18, 2020 at 16:01 | comment | added | Angelos Amyntas | The ziformula is a binomial model turned up side down, it tells you which of your predictors increase the probability of no drinks vs some drinks. The conditional model tells you which predictors influence the number of drinks, for those cases where some drinks have been had. (I am now really curious about your predictors) | |
| May 18, 2020 at 14:00 | comment | added | alessothegreat | Okay it looks like the conditional model is predicting 0 or the "absence" of the variable you are looking at. | |
| May 18, 2020 at 13:15 | comment | added | alessothegreat | Thanks, that is a helpful reframe of what is happening, I had trouble interpreting that part of the document. And I've updated my model to include the random effects as well. So, if the zero-inflation is predicting likelihood of 0 vs. not-zero, is the conditional model just predicting 0 vs. 1 like a true logistic regression? | |
| May 17, 2020 at 8:22 | history | edited | Angelos Amyntas | CC BY-SA 4.0 |
added 217 characters in body
|
| May 17, 2020 at 8:05 | history | answered | Angelos Amyntas | CC BY-SA 4.0 |