compute_c_inv: compute_c_inv computes the uncertainty of the random effects...

In L-Ippel/SEMA: Streaming Expectation Maximization Approximation

Description

compute_c_inv computes the uncertainty of the random effects coefficients. NOTE, this function contains a matrix inversion of the variance of the random effects, when this matrix is large, the SEMA algorithm slows down from raudenbush and bryk (2002) Hierarchial linear models, 2nd edition, EQUATION 14.15b

Usage

compute_c_inv(z_sq, resid_var, random_var)

Arguments

z_sq	Square matrix of the squared random effects variables.
resid_var	Residual variance, a scalar.
random_var	Variance of the random effects, a matrix.

Value

Inverted matrix with updated uncertainty of the random effects coefficients.

L-Ippel/SEMA documentation built on May 30, 2019, 8:23 a.m.