balance (version 0.2.4)

sbp.fromABA: Build SBP Matrix of "Anti-Principal Balances"

Description

This function builds an SBP of "anti-principal balances" by clustering the difference of the log-ratio variance from the maximum log-ratio variance. Unlike principal balances, where the first balances explain the most variance, this function selects "anti-principal balances" so that the last balances explain relatively more variance.

Usage

sbp.fromABA(x, alpha = NA)

Arguments

x

A matrix with rows as samples (N) and columns as components (D).

alpha

A double. Defines a hyper-parameter used by the Box-Cox transformation to approximate log-ratio variance in the presence of zeros. Skip with NA.

Value

An SBP matrix.

Examples

# NOT RUN {
library(balance)
data(iris)
x <- iris[,1:4]
sbp.fromABA(x)

# }