update_a_ic: Update one element of scores A

Description Usage Arguments Value

View source: R/exp_pca.R

Description

Update one element of scores A

Usage

1
update_a_ic(obj_fun, obj_grad_a, x_i, a_ic, v_c, s_i, lambda, mu0)

Arguments

obj_fun

Version of loss given by bregman divergence between data and link to use in current updates.

obj_grad_a

Gradient of the provided loss, with respect to a.

x_i

The i^th sample of X.

a_ic

The initial value of a_ic to start the search from.

v_c

The c^th component of V.

s_i

The s^th row of the matrix AV^T - a_cv_c^T

lambda

The regularization parameter in the optimizatyion.

mu0

The value to regularize towards.

Value

a_ic The updated version of a_ic.


krisrs1128/expPCA documentation built on May 18, 2017, 11:45 p.m.