Skip to contents

Estimate ILR Parameters from Compositional Samples

gc_ilr_params.Rd

Convert compositional samples to Isometric Log-Ratio (ILR) space and compute the mean vector and covariance matrix. These statistics are used to parameterize the multivariate geostatistical model.

Usage

gc_ilr_params(samples)

Arguments

samples

A data frame of compositions (each row sums to constant, typically 100% or 1).

Value

A list containing:

  • mean: Vector of ILR means (length D-1 where D is number of components)

  • cov: Covariance matrix of ILR values (dimensions (D-1) x (D-1))

  • names: Original component names (character vector)

  • base_class: Class used for transformation ("acomp")

Details

This function:

  1. Stores original column names for later use

  2. Converts to compositions::acomp (absolute composition)

  3. Applies compositions::ilr transformation

  4. Computes column means and covariance matrix of ILR values

The ILR transformation eliminates the sum constraint, allowing standard multivariate geostatistics to be applied.

Examples

# Simulate some simple compositions
samples <- data.frame(
  sand = c(20, 25, 30, 22),
  silt = c(60, 55, 50, 58),
  clay = c(20, 20, 20, 20)
)

params <- estimate_ilr_params(samples)
#> Error in estimate_ilr_params(samples): could not find function "estimate_ilr_params"
str(params)
#> Error: object 'params' not found
print(params$mean)
#> Error: object 'params' not found
print(params$cov)
#> Error: object 'params' not found