mleLR {countprop}

Maximum Likelihood Estimate for multinomial logit-normal model

Description

Returns the maximum likelihood estimates of multinomial logit-normal model parameters given a count-compositional dataset. The MLE procedure is based on the multinomial logit-Normal distribution, using the EM algorithm from Hoff (2003).

Usage

mleLR(
  y,
  max.iter = 10000,
  max.iter.nr = 100,
  tol = 1e-06,
  tol.nr = 1e-06,
  lambda.gl = 0,
  gamma = 0.1,
  verbose = FALSE
)

Arguments

y

Matrix of counts; samples are rows and features are columns.

max.iter

Maximum number of iterations

max.iter.nr

Maximum number of Newton-Raphson iterations

tol

Stopping rule

tol.nr

Stopping rule for the Newton-Raphson algorithm

lambda.gl

Penalization parameter lambda, for the graphical lasso penalty. Controls the sparsity of Sigma

gamma

Gamma value for EBIC calculation of the log-likelihood

verbose

If TRUE, print information as the functions run

Value

The additive log-ratio of y (v); maximum likelihood estimates of mu, Sigma, and Sigma.inv; the log-likelihood (log.lik); the EBIC (extended Bayesian information criterion) of the log-likelihood of the multinomial logit-Normal model with the graphical lasso penalty (ebic); degrees of freedom of the Sigma.inv matrix (df).

Note

The graphical lasso penalty is the sum of the absolute value of the elements of the covariance matrix Sigma. The penalization parameter lambda controls the sparsity of Sigma.

This function is also used within the mlePath() function.

Examples

data(singlecell)
mle <- mleLR(singlecell)

mle$mu
mle$Sigma
mle$ebic