prep_predata: Creation of predata and data impuation of missing data.

In MathieuGenu/geffaeR: Generic Function for Abundance Estimation with R

Creation of predata and data impuation of missing data.

Description

Creation of predata, a data.frame containing environmental parameters in a grid of the study area. Also it allows to impute missing data for predata and segdata using PCA from the missMDA package or Amelia from Amelia package.

Usage

prep_predata(
  segdata,
  gridfile_name,
  varenviro,
  do_log_enviro,
  varphysio,
  do_log_physio,
  imputation = "Amelia",
  shape,
  saturate_predata = F,
  saturate_segdata = F,
  inbox_poly = T,
  col2keep = NULL,
  verbose = TRUE
)

Arguments

segdata	"Segdata" data.frame, built with prepare_data_effort.
gridfile_name	Grid of the study area with all the environmental variables which allows to build predata.
varenviro	Vector containing all dynamic environmental variables (e.g. chlorophyl, SST).
do_log_enviro	Vector precising which variables among varenviro that are necessary to transform to neperian logarithm + 1 (log1p).
varphysio	Vector containing non-dynamic environmental variables (i.e depth, dist200m).
do_log_physio	Vector precising which variables among varphysio that are necessary to transform to neperian logarithm + 1 (log1p).
imputation	Data imputation method to missing values of segdata and predata. 2 methods allowed : "PCA" : PCA method from missMDA. "Amelia" : Amelia method from Amelia. Default method is PCA.
shape	Shapefile of the study area. It can be either a SpatialPolygonsDataFrame class object, in this case it is not necessary to give shape_layer argument. Or it can be the name of the shape object with its extension ".shp" (ex : "data/studyAreaShapefile.shp").
saturate_predata	Boolean. If TRUE, saturate function is applied on all varenviro an varphysio columns of segdata. Saturate function excludes extreme valuesand keep values between quantiles 95 percent and 5 percent.
saturate_segdata	Boolean. If TRUE, saturate function is applied on all varenviro an varphysio columns of predata. Saturate function excludes extreme valuesand keep values between quantiles 95 percent and 5 percent.
inbox_poly	When TRUE, keep only part of the grid that match with the study area.
col2keep	character string corresponding to the columns wanted to appear in output in predata.
verbose	Boolean. whether or not to display the progress bar for "PCA" or iteration number for "amelia".

Value

This function return a list containing :

1. predata : data.frame of predata. Corresponding to gridata keeping covariates, coordinates and area of cells

2. segdata : data.frame of segdata.

3. pca_seg : Output of PCA function on predata.

4. pca_pred : Output of PCA function on segdata.

5. seg_mipat : output of VIM::aggr on predata

6. pred_mipat : output of VIM::aggr on segdata

MathieuGenu/geffaeR documentation built on March 23, 2022, 7:50 p.m.