R: Fit models with TMB

glmmTMB {glmmTMB}

R Documentation

Fit models with TMB

Description

Fit models with TMB

Usage

glmmTMB(formula, data = NULL, family = gaussian(), ziformula = ~0,
  dispformula = ~1, weights = NULL, offset = NULL, se = TRUE,
  verbose = FALSE, doFit = TRUE, control = glmmTMBControl())

Arguments

`formula`	combined fixed and random effects formula, following lme4 syntax
`data`	data frame
`family`	family (variance/link function) information; see `family` for generic family details or `family_glmmTMB` for details of `glmmTMB` specific families. As in `glm`, `family` can be specified as (1) a character string referencing an existing family-construction function (e.g. ‘"binomial"’); (2) a symbol referencing such a function (‘binomial’); or (3) the output of such a function (‘binomial()’). In addition, for families such as `betabinomial` that are special to `glmmTMB`, family can be specified as (4) a list comprising the name of the distribution and the link function (‘list(family="binomial", link="logit")’). However, the first 3 options are preferable.
`ziformula`	a one-sided (i.e., no response variable) formula for zero-inflation combining fixed and random effects: the default `~0` specifies no zero-inflation. Specifying `~.` will set the right-hand side of the zero-inflation formula identical to the right-hand side of the main (conditional effects) formula; terms can also be added or subtracted. Offset terms will automatically be dropped from the conditional effects formula. The zero-inflation model uses a logit link.
`dispformula`	a one-sided formula for dispersion containing only fixed effects: the default `~1` specifies the standard dispersion given any family. The argument is ignored for families that do not have a dispersion parameter. For an explanation of the dispersion parameter for each family, see (`sigma`). The dispersion model uses a log link. In Gaussian mixed models, `dispformula=~0` fixes the parameter to be 0, forcing variance into the random effects.
`weights`	weights, as in `glm`. Not automatically scaled to have sum 1.
`offset`	offset
`se`	whether to return standard errors
`verbose`	logical indicating if some progress indication should be printed to the console.
`doFit`	whether to fit the full model, or (if FALSE) return the preprocessed data and parameter objects, without fitting the model
`control`	control parameters; see `glmmTMBControl`.

Details

binomial models with more than one trial (i.e., not binary/Bernoulli) can either be specified in the form prob ~ ..., weights = N or in the more typical two-column matrix (cbind(successes,failures)~...) form.
in all cases glmmTMB returns maximum likelihood estimates - random effects variance-covariance matrices are not REML (so use REML=FALSE when comparing with lme4::lmer), and residual standard deviations (sigma) are not bias-corrected. Because the df.residual method for glmmTMB currently counts the dispersion parameter, one would need to multiply by sqrt(nobs(fit)/(1+df.residual(fit))) when comparing with lm ...
by default, vector-valued random effects are fitted with unstructured (general positive definite) variance-covariance matrices. Structured variance-covariance matrices can be specified in the form struc(terms|group), where struc is one of
- diag (diagonal, heterogeneous variance)
- ar1 (autoregressive order-1, homogeneous variance)
- cs (compound symmetric, heterogeneous variance)
- ou (* Ornstein-Uhlenbeck, homogeneous variance)
- exp (* exponential autocorrelation)
- gau (* Gaussian autocorrelation)
- mat (* Matérn process correlation)
- toep (* Toeplitz)
(note structures marked with * are experimental/untested)

Examples

(m1 <- glmmTMB(count~ mined + (1|site), 
  zi=~mined, 
  family=poisson, data=Salamanders))
summary(m1)

## Zero-inflated negative binomial model
(m2 <- glmmTMB(count~spp + mined + (1|site), 
  zi=~spp + mined, 
  family=nbinom2, Salamanders))

## Hurdle Poisson model
(m3 <- glmmTMB(count~spp + mined + (1|site), 
  zi=~spp + mined, 
  family=truncated_poisson, Salamanders))

## Binomial model
data(cbpp, package="lme4")
(tmbm1 <- glmmTMB(cbind(incidence, size-incidence) ~ period + (1 | herd),
               data=cbpp, family=binomial))

## Dispersion model
sim1=function(nfac=40, nt=100, facsd=.1, tsd=.15, mu=0, residsd=1)
{
  dat=expand.grid(fac=factor(letters[1:nfac]), t= 1:nt)
  n=nrow(dat)
  dat$REfac=rnorm(nfac, sd= facsd)[dat$fac]
  dat$REt=rnorm(nt, sd= tsd)[dat$t]
  dat$x=rnorm(n, mean=mu, sd=residsd) + dat$REfac + dat$REt
  return(dat)
}
set.seed(101)
d1 = sim1(mu=100, residsd =10)
d2 = sim1(mu=200, residsd =5)
d1$sd="ten"
d2$sd="five"
dat = rbind(d1, d2)
m0 = glmmTMB(x~sd+(1|t), dispformula=~sd, dat)
fixef(m0)$disp
c(log(5^2), log(10^2)-log(5^2)) #expected dispersion model coefficients

[Package glmmTMB version 0.2.0 Index]