Plotting function to assess the mean-variance relationship of generated data

Usage

plotVariance(
  model,
  simdat,
  cuts = 20,
  quants = c(0.025, 0.975),
  n.sim = NULL,
  store.data = FALSE
)

Arguments

model: model object of class glm, gam or gamMRSea
simdat: matrix of simulated response data. Each column is a new simulated data set
cuts: The number of cut points for calculation of residual variance
quants: vector of length 2 stating the quantiles for the confidence interval bands for the simulated mean-variance relationship
n.sim: (default = NULL) denotes how many simulations from simdat to use. If NULL then all are used.
store.data: Logical indicating whether or not to return the plotting data. Default is FALSE (no data stored).

Author

Lindesay Scott-Hayward (University of St Andrews)

Examples

data(nystedA_slim)

initialModel<-MRSea::gamMRSea(response ~ 1 + as.factor(yearmonth)+depth +
                     x.pos + y.pos + offset(log(area)),  data=nysted,
                     family=quasipoisson)

nsim<-550
d<-as.numeric(summary(initialModel)$dispersion)
newdat<-generateNoise(nsim, fitted(initialModel), family='poisson', d=d)

plotVariance(initialModel, newdat)
#> Warning: Simulations:  2  removed from plotting owing to extra ordinary values. The vector of ID's has been put in the workspace: badsim.id
#> Warning: The `<scale>` argument of `guides()` cannot be `FALSE`. Use "none" instead as
#> of ggplot2 3.3.4.
#> ℹ The deprecated feature was likely used in the MRSeaPower package.
#>   Please report the issue to the authors.