galamm (version 0.3.0)

residuals.galamm: Residuals of galamm objects

Description

Computes residuals for models fit with galamm() using the definitions in Chapter 8 of dunnGeneralizedLinearModels2018;textualgalamm. Define \(y\) as the response and \(\hat{\mu}\) as the model fit. Importantly, \(\hat{\mu}\) includes all random effects. Also define \(V(\cdot)\) as the variance function of the model family, and \(w\) as the weight. The Pearson residual is then $$r_{P} = (y - \hat{\mu})/\sqrt{V(\hat{\mu}) / w}.$$ Furthermore, let \(sgn(\cdot)\) be the function which returns the sign of its argument and let \(d(y, \hat{\mu})\) be the model deviance. The deviance residual is then $$r_{D} = sgn(y - \hat{\mu}) \sqrt{w d(y, \hat{\mu})}.$$

Usage

# S3 method for galamm
residuals(object, type = c("pearson", "deviance"), ...)

Value

Numeric vector of residual values.

Arguments

object

An object of class galamm returned from galamm.

type

Character of length one describing the type of residuals to be returned. One of "pearson" and "deviance". Argument is case sensitive.

...

Optional arguments passed on to other methods. Currently not used.

References

See Also

fitted.galamm() for model fitted values, predict.galamm() for model predictions, and plot.galamm() for diagnostic plots. The generic function is residuals().

Other details of model fit: VarCorr(), coef.galamm(), confint.galamm(), deviance.galamm(), factor_loadings.galamm(), family.galamm(), fitted.galamm(), fixef(), formula.galamm(), llikAIC(), logLik.galamm(), nobs.galamm(), predict.galamm(), print.VarCorr.galamm(), ranef.galamm(), response(), sigma.galamm(), vcov.galamm()

Examples

# Poisson GLMM
count_mod <- galamm(
  formula = y ~ lbas * treat + lage + v4 + (1 | subj),
  data = epilep, family = poisson
)

# Extract residuals
residuals(count_mod)