fgem_grad: Calculate fgem gradient

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

Description

Calculate fgem gradient

Usage

fgem_grad(par, X, BF, l2 = 0, l1 = 0, log_BF = FALSE, ...)

## S3 method for class 'matrix'
fgem_grad(par, X, BF, l2 = 0, l1 = 0, log_BF = FALSE, ...)

## S3 method for class 'dgCMatrix'
fgem_grad(par, X, BF, l2 = 0, l1 = 0, log_BF = FALSE, ...)

## Default S3 method:
fgem_grad(par, X, BF, l2 = 0, l1 = 0, log_BF = FALSE, ...)

Arguments

par	Beta parameter
X	matrix (can be Matrix from Matrix package)
BF	bayes factor (or log Bayes factor if log_BF=TRUE)
l2	L2 regularization
log_BF	boolean indicating whether BF should be treated as log_BF

Value

vector with gradients

Methods (by class)

• matrix: gradient for dense matrix

• dgCMatrix: gradient for sparse matrix

• default: gradient for generic

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