Fit a multivariate normal model without covariates or covariance restrictions. In addition to the (straightforward) parameter estimates the fitted log-likelihood and corresponding score contributions are computed.

mvnfit(
  y,
  x = NULL,
  start = NULL,
  weights = NULL,
  offset = NULL,
  model = c("correlation", "mean", "variance"),
  ...,
  estfun = FALSE,
  object = FALSE
)

Arguments

y

A matrix or data.frame where each row corresponds to a k-dim observation.

x

Not used yet

start

Not used yet

weights

Not used yet

offset

Not used yet

model

Vector of characters. Specifies which estimated parameters are returned.

...

Not used yet

estfun

Logical. Should the matrix of score contributions (aka estimating functions) be returned?

object

Not used yet

Details

Used internally in when method="mob"