apply_ordination: Apply ordination procedures for multivariate statistical...

Description Usage Arguments Value

View source: R/apply_ordination.R

Description

#' Applies a given ordination procedure to a data set and returns a ordination object.

Usage

1
2
3
4
apply_ordination(data, protocol = "1", dimensions = 2,
  exception_columns = NULL, variable_tags = NULL, coda_override = NULL,
  coda_transformation_method = "CLR", coda_alr_base = 1,
  coda_pca_method = "robust", init_seed = 0, coda_samples = 100)

Arguments

data

Data frame, including compositional and petrographic data.

protocol

Character, cerUB protocol to be applied. "1": Analysis of compositional data; "2a": Analysis of petrographic data (relative ranking difference); "2b": Analysis of petrographic data (neighbor interchange); "3": Analysis of compositional data and petrographic data (relative ranking); "4": Analysis of compositional data and petrographic data (relative ranking) to characterize provenance.

dimensions

Numeric, number of dimensions of the ordination object.

exception_columns

Numeric, the vector of variables names to be searched for exceptions.

variable_tags

Character, two-column data frame containing (1) the names of variables and (2) their tags.

coda_override

Character, vector with the names of the compositional variables.

coda_transformation_method

Character, the log-ratio transformation to be applied: "ALR" for additive log-ratio, "CLR" for centered log-ratio, "ILR" for isometric log-ratio. Additionally, accepts "log" for applying logarithmic transformation and "std" for standardization (scaled and centred).

coda_alr_base

Character/Numeric, the name/index of the variable to be used as divisor in additional log-ratio transformation.

coda_pca_method

Character, Principal Components Analysis (PCA) method: "standard" for standard PCA, "robust" for robust PCA.

init_seed, coda_samples

Numeric, arguments passed to princomp_coda.

Value

Ordination object containing the projection of observations (scores) and variables (loadings) in 'n' dimensions, the distance matrix used (dist_matrix), and an approximation of the fitness of projections.


Andros-Spica/cerUB documentation built on June 9, 2020, 9:22 p.m.