zVarArrayError: Variation array relative error

In Japal/zCompositions: Treatment of Zeros, Left-Censored and Missing Values in Compositional Data Sets

Variation array relative error

Description

This function computes squared relative errors of variation arrays per group with respect to the overall variation array based on observed data in a compositional data set. Groups can be defined by either zero/unobserved data patterns or by a grouping factor in fully observed zero-free data sets.

Usage

zVarArrayError(
  X,
  label = 0,
  groups = NULL,
  breakdown = FALSE,
  suppress.print = FALSE
)

Arguments

X	Compositional data set (matrix or data.frame class).
label	Unique label (numeric or character) used to denote zeros/unobserved data in X (label = 0, default).
groups	Grouping factor in fully observed zero-free data sets (groups = NULL, default).
breakdown	Logical value. Show results broken down by group (breakdown = FALSE, default).
suppress.print	Suppress printed feedback (suppress.print = FALSE, default).

Details

Squared relative errors (SRE) are calculated by confronting variation arrays (log-ratio variances and means) obtained per group and the overall variation array based on observed data. Raw SREs are computed for each available pair-wise log-ratio. The weighted version uses the corresponding group sizes to weight raw SREs. Total SRE is obtained as the sum of weighted SREs for each log-ratio. Further details by group are provided by setting breakdown = TRUE.

Value

1. SRE for each log-ratio variance and mean. 2. Weighted SRE for each log-ratio variance and mean. 3. Total SRE across log-ratio variances and means. 4. Percentage contribution of each log-ratio to SRE in log-ratio variances and means. If breakdown = TRUE: 4. SREs per group. 5. Weighted SREs per group. 6. Percentage contribution of each group to total SRE.

See Also

zPatterns, zVarArray.

Examples

data(Water)
zPatterns(Water, label = 0)
zVarArrayError(Water)
zVarArrayError(Water, breakdown = TRUE)

# From a completed data set

data(mdl) # matrix of limits of detection for Water
Water_multKM <- multKM(Water,label=0,dl=mdl) # nondetects imputation

# Results split by two ficticious groups A and B
zVarArrayError(Water_multKM,groups=rep(c("A","B"),each=50))

Japal/zCompositions documentation built on March 17, 2024, 5:19 p.m.