cv_relax_fgem: fit cross-validated FGEM model and then fit a relaxed model...

In CreRecombinase/FGEM: R package for Functional Genomics using Expectation Maximization (FGEM)

Description

fit cross-validated FGEM model and then fit a relaxed model (no L1 penalty) on the best

Usage

cv_relax_fgem(
  X,
  BF,
  log_BF = FALSE,
  alpha = 0.95,
  nlambda = 100,
  lambda = NULL,
  stratify_BF = TRUE,
  v = 10,
  ...
)

Arguments

X	Feature matrix
BF	vector of bayes factors
alpha	alpha (as in glmnet). Alpha is the (scalar) proportion of 'lambda' applied to l1 penalization, while '1-alpha' is applied to l2
lambda	vector of shrinkage parameters

CreRecombinase/FGEM documentation built on July 17, 2020, 5:30 a.m.