Skip to contents

In the Model 2a factor parameterization, the same-parameter cross-biomarker covariance for biomarkers 1 and 2 is the product of their loadings: $$c_p = \lambda_{1,p}\,\lambda_{2,p}.$$ This pure helper computes c_p for one MCMC draw (or any single set of loadings). It contains no fitting logic and is trivial to unit-test.

Usage

cross_cov_from_loadings(lambda_mat)

Arguments

lambda_mat

A K x P matrix of loadings, rows = biomarkers, columns = kinetic parameters. Only the first two rows are used (Model 2a pairs two biomarkers).

Value

A length-P numeric vector of cross-biomarker covariances.