Capturing between-tasks covariance and similarities using multivariate linear mixed models (Q2209832)

The authors present the MrRCE method to produce an estimator of the covariance components and a predictor of the multivariate regression coefficient matrix. Their approach exploits similarities among random coefficients and accounts for correlated errors. They propose an efficient EM-based algorithm for computing the MrRCE. Using simulated and real data, they illustrate that the proposed method can outperform the commonly used methods for multivariate regression, in settings were errors or coefficients are related. The paper is dealing with a very interesting mathematical problem.