sup_class: Function for eigenanalysis and supervised classification
In phenofun: Automated Processing of Webcam Images for Phenological Classification

Description Usage Arguments Value

This function allows to carry out an eigenanalysis of given webcam data and a supervised classification of DOYs with respect to seasons. Example code can be downloaded at http://bothmann.userweb.mwn.de/dissertation.html.

sup_class(folder_results = paste0(getwd(), "/"),
  folder_data = paste0(getwd(), "/"),
  name_of_analysis = substr(as.character(Sys.time()), 1, 10),
  lists_files = lists_files, do_eigenanalysis = TRUE,
  do_classification = TRUE, do_classification_newyear = FALSE,
  do_masking = FALSE, load_eigen = FALSE, show_eigen = FALSE,
  varimax_rotation = FALSE, sum1 = FALSE, norm_it = TRUE,
  norm_type = "mean0", colorspace = "rgb", total_var = TRUE,
  col_in = 1:3, colormode = "Color", filename_mask = NULL,
  pfad_eigen = NULL, fact = NULL, x = NULL, y = NULL, training_year,
  test_year, season_list, n_pc = 12, year, doy, complete_year = TRUE,
  n_images_try = 10, seasons_newyear, mask = NULL)

`folder_results`	Folder where the results will be saved
`folder_data`	Folder where the data is saved
`name_of_analysis`	Name of the specific analysis for a unique folder
`lists_files`	Vector of file names of images
`do_eigenanalysis`	If `TRUE` (default), eigenanalysis is carried out
`do_classification`	If `TRUE` (default), supervised classification is carried out
`do_classification_newyear`	If `TRUE`, supervised classification of a new/test year is carried out, default is `FALSE`
`do_masking`	If `TRUE`, images are masked with a given ROI as first step is carried out, default is `FALSE`
`load_eigen`	If `TRUE`, results of eigenanalysis are loaded, default is `FALSE`
`show_eigen`	If `TRUE`, eigenimages are save, default is `FALSE`
`varimax_rotation`	If `TRUE`, principal components are rotated with varimax method, default is `FALSE`
`sum1`	If `TRUE`, images are normed to sum=1, default is `FALSE`
`norm_it`	If `TRUE` (default), images are normed before the analysis
`norm_type`	Type of norming
`colorspace`	Color space (rgb, hsv, etc.)
`total_var`	If `TRUE` (default), total variance is computed in eigenanalysis
`col_in`	Color channels to be read in, default `1:3` corresponds to red, blue and green
`colormode`	The color mode
`filename_mask`	File name of mask / ROI
`pfad_eigen`	Path of results of eigenanalysis to be loaded
`fact`	Weighting factor for sum and diff of mean and eigenimages, default is `NULL` and then quantiles of the scores are used as weights
`x`	vector of x-coordinates of analyzed pixels
`y`	vector of y-coordinates of analyzed pixels
`training_year`	The training year
`test_year`	The test year
`season_list`	List specifying the true seasons
`n_pc`	Number of principal components / eigenimages
`year`	Vector of year for each image
`doy`	Vector of DOYs for each image
`complete_year`	If `TRUE` (default), the complete year is analysed, else the first `n_images_try` are analyzed (only for testing)
`n_images_try`	See complete_year
`seasons_newyear`	Vector with true seasons of the test year
`mask`	Do not edit, mask is loaded, if needed