logitNormalVariation: Logit Normal Variation

In countprop: Calculate Model-Based Metrics of Proportionality on Count-Based Compositional Data

Logit Normal Variation

Description

Estimates the variation matrix of count-compositional data based on a multinomial logit-Normal distribution. Estimation is performed using only the parameters of the distribution.

Usage

logitNormalVariation(
  mu,
  Sigma,
  type = c("standard", "phi", "phis", "rho"),
  order = c("second", "first")
)

Arguments

mu	The mle estimate of the mu matrix
Sigma	The mle estimate of the Sigma matrix
type	Type of variation metric to be calculated: standard, phi, phis (a symmetrical version of phi), rho, or logp (the variance-covariance matrix of log-transformed proportions)
order	The order of the Taylor-series approximation to be used in the estimation

Value

An estimate of the requested metric of proportionality.

Examples

data(singlecell)
mle <- mleLR(singlecell)
mu.hat <- mle$mu
Sigma.hat <- mle$Sigma

logitNormalVariation(mu.hat, Sigma.hat)
logitNormalVariation(mu.hat, Sigma.hat, type="phi")
logitNormalVariation(mu.hat, Sigma.hat, type="rho")

countprop documentation built on Aug. 18, 2023, 9:06 a.m.