Modeling zero-inflated count data with glmmTMB
read more
Citations
The coefficient of determination R2 and intra-class correlation coefficient from generalized linear mixed-effects models revisited and expanded.
The coefficient of determination R2 and intra-class correlation coefficient from generalized linear mixed-effects models revisited and expanded
Accounting for individual-specific variation in habitat-selection studies: Efficient estimation of mixed-effects models using Bayesian or frequentist computation
Violating the normality assumption may be the lesser of two evils
Violating the normality assumption may be the lesser of two evils
References
R: A language and environment for statistical computing.
Fitting Linear Mixed-Effects Models Using lme4
Regularization Paths for Generalized Linear Models via Coordinate Descent
Mixed Effects Models and Extensions in Ecology with R
Generalized linear mixed models: a practical guide for ecology and evolution
Related Papers (5)
Frequently Asked Questions (7)
Q2. How are the counts of owls measured?
Ecological phenomena are often measured in the form of discrete count data, e.g., the number of times that owl nestlings beg for food (Roulin & Bersier, 2007), counts of salamanders in streams (Price et al., 2016), or counts of parasite eggs in fecal samples of sheep (Brown et al., 2012).
Q3. How long did it take to fit the model to the original data?
185Fitting the model to the original data replicated to have more observations per site required, on average, half the time to fit with INLA compared to glmmTMB, 22 times as long with glmmADMB, 30 times as long with lme4, and 59 times as long with brms (Fig A.2).
Q4. Why are pscl and mgcv not included in the glmm?
lme4 and mgcv are not included because65 they can only estimate zero-inflation when wrapped in an iterative algorithm (Minami et al., 2007; Bolker et al., 2013).
Q5. What is the common use of the statistical computing environment R?
Many biologists use the statistical computing environment R and its con-tributed packages to organize, model, and graph their data (R Core Develop-35ment Team, 2016).
Q6. How can pscl be used to test the hypothesis that sheep fecal?
40For example, pscl can be used to test the hypothesis that sheep fecal egg counts depend on age and extra zeros depend on genotype.
Q7. How fast was the fit to simulated data?
Fitting this model to simulated data with the same structure as the original data was, on average, equally fast in glmmTMB and INLA, 26 times slower with glmmADMB, 30 times slower with lme4, and 274 times slower with brms (Fig A.1).