eempf_residuals_plot {staRdom}

Plot samples by means of whole sample, each single component and residuum

Description

A raster of plots is created. Each column shows one sample. The top n rows show the n components from the model according their occurance in the certain samples. The second last row shows the residual, not covered by any component in the model and the last row shows the whole sample.

Usage

eempf_residuals_plot(
  pfmodel,
  eem_list,
  res_data = NULL,
  spp = 5,
  select = NULL,
  residuals_only = FALSE,
  cores = parallel::detectCores(logical = FALSE),
  contour = FALSE,
  colpal = "default"
)

Arguments

pfmodel

object of class parafac containing the generated model

eem_list

object of class eemlist with all the samples that should be plotted

res_data

optional, data of sample residuals related to the model, output from eempf_residuals

spp

optional, samples per plot

select

optional, character vector of samples you want to plot

residuals_only

plot only residuals

cores

number of cores to use for parallel processing

contour

logical, states whether contours should be plotted

colpal

"default" to use the viridis colour palette, "rainbow" to use a subset of the rainbow palette, any custom vector of colors or a colour palette. A gradient will be produced from this vector. Larger vectors (e.g. 50 elements) can produce smoother gradients.

Details

eem_list may contain samples not used for modelling. Calculation is done by A_missing. This especially interesting if outliers are excluded prior modelling and should be evaluated again afterwards. Usually, residuals contain negative values, while these is the exception in samples and PARAFAC components. Therefore, we decided to use a similar colour palette as in the other plot functions but adding a purple tone for negative values.

Value

several ggplot objects

Examples

data(eem_list)
data(pf_models)

eem_list <- eem_extract(eem_list, 1:10)

eem_list <- eem_rem_scat(eem_list, rep(TRUE, 4), c(15,10,16,12))

eempf_residuals_plot(pf4[[1]], eem_list, cores = 2)

# use other colour schemes:
# eempf_residuals_plot(pf4[[1]], eem_list, colpal = c("blue",heat.colors(50)))
# plots <- eempf_residuals_plot(pf4[[1]], eem_list)
# lapply(plots, function(pl){
#   pl +
#     scale_fill_viridis_c() +
#     scale_colour_viridis_c()
# })