Function to run a power analysis for a model of type 'gamMRSea'

Usage

powerSimPll(
  newdat,
  model,
  empdistribution,
  nsim,
  powercoefid,
  predictionGrid = NULL,
  g2k = NULL,
  splineParams = NULL,
  n.boot = 500,
  nCores = 1
)

Arguments

newdat: matrix of values whose rows are observations and columns are simulated sets of response data
model: model object of class gamMRSea. This is the model used to generate the data, updated for the inclusion of an event term.
empdistribution: Empirical distribution for the runs test
nsim: number of simulations to compute
powercoefid: the ID of the coefficient in model that represents the event term
predictionGrid: dataframe containing a coordinate grid with covariates. This grid is used in the assessment of before and after change differences.
g2k: object required for MRSea model with spline terms
splineParams: object required for MRSea model with spline terms
n.boot: Number of bootstraps to use to get percentile based confidence intervals for predictions.
nCores: How many cores to use for parallel processing. If one, then parallel computing is not used and a progress bar is printed. If parallel, estimated time till finish along with a progress bar is provided.

Value

An object of class gamMRSea.power.

rawrob: Vector of zeros and ones indicating whether raw or robust s.e. have been used.
imppvals: p-values for the event change term for each simulated run
betacis: Confidence intervals for the event term coefficient (only if site-wide change)
powsimfits: Fitted values each model
bootdifferences: data frame of bootstrapped differences (before and after event). Number of rows is the number of rows in the prediction grid divided by 2
bootpreds: data frame of bootstrapped predictions.
Abundance: Table of sitewide abundance (with upper and lower 95\% CI)
link: List of the link function (link) and inverse link function (linkinv)

Author

LAS Scott-Hayward, University of St Andrews

Examples

# See the vignette for an example of how to run this code.