var.ricf: Maximum Likelihood Method for Linear Structural Equation...

In ysamwang/BCD: Linear Structural Equations with Gaussian Errors

Maximum Likelihood Method for Linear Structural Equation Models

Description

Calculate Fisher information for non-structural zeros in B and Omega matrix. The ordering of the estimates in the Information matrix follows standard vec(-) convention indexing over rows first then columns. The elements of B are ordered before the elements of Omega. Duplicate elements of Omega are not included.

Usage

var.ricf(Y, B, Omega, B.hat, Omega.hat, type = "expected")

Arguments

Y	V by n data matrix where each row corresponds to an observed variable and each column corresponds to a multivariate observation
B	V by V matrix with 0,1 giving structure of directed edges
Omega	V by V matrix with 0,1 giving structure of bi-directed edges
B.hat	V by V matrix giving estimated edges weights for directed edges
Omega.hat	V by V matrix giving estimated edge weights for bi-directed edges
type	string describing which Fisher Information to calculate. Options are "expected", "observed", or "sandwich".

Value

The inverse (scaled by n) Fisher information matrix as derived by Fox and Drton 2014 or the Huber-White misspecified model covariance estimate

ysamwang/BCD documentation built on Sept. 3, 2023, 1:33 a.m.