High Dimension Prediction

Prediction methods may be high dimension in either outcome or covariate matrix. Here we examine shrinkage methods for multivariate prediction, multivariate calibration, and penalized regression for a single outcome with a large number of covariates--we motivate this section with microbiome data.