MCAP is a new multilevel method for regressing covariance matrices on covariates that models cluster-specific projections on the unit sphere with a von Mises-Fisher distribution and estimates parameters via hierarchical likelihood maximization.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
stat.ME 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Multilevel Regression Modeling of Covariance Matrix Outcomes
MCAP is a new multilevel method for regressing covariance matrices on covariates that models cluster-specific projections on the unit sphere with a von Mises-Fisher distribution and estimates parameters via hierarchical likelihood maximization.