Table 2.
Results from the GLM regression.
| Model | β0 | β1 | β2 | α | ΔELPD | |
|---|---|---|---|---|---|---|
| Priors | 𝒩(0, 102) | HalfCauchy(1) | ||||
| NB | (Eq. 1) | ![]() |
![]() |
![]() |
0 | |
| NB with B/T dependence | (Eq. 2) | ![]() |
![]() |
![]() |
![]() |
0.25 |
| Poisson with B/T dependence | (Eq. 2) | ![]() |
![]() |
![]() |
45.06 | |
| Poisson | (Eq. 1) | ![]() |
![]() |
46.76 | ||
Notes. Two first rows using the negative binomial likelihood, while the third and fourth ones use Poisson likelihood. Each column indicates one of the parameters of the fit. The over dispersion parameter of the negative binomial regression is α. Models are sorted by the expected log pointwise predictive density (ELPD). Poisson regression can be clearly ruled out. Both negative binomial regression models have similar score according to ELPD, moreover, the interaction term (β2) is compatible with 0 within the 95% high density interval of the posterior.
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.











