From: Bayesian negative binomial logit hurdle and zero-inflated model for characterizing smoking intensity
Likelihood ratio Test for nested models | ||||
---|---|---|---|---|
Model | Hypothesis | p-value | Preferable model | |
ZIP = PLH vs ZINB = NBLH | H0:\(\varnothing \) =0 VS H1:\(\varnothing \) >0 |  < 2.2e−16 | ZINB = NBLH |
Vuong test for non-nested models | ||||
---|---|---|---|---|
Model | Hypothesis | Vuong statistics | p-value | Comment |
ZIP vs PLH | Model1 = model2 vs Model1 > model2 | 1.216179 | 0.11196 | ZIP = PLH |
ZINB vs NBLH | Model1 = model2 vs Model1 > model2 | 1.234438 | 0.10852 | ZINB = NBLH |