apg_progect: Accelerated Proximal Gradient based projection method

In kcygan/dryclean: rPCA based method to de-noise genomic coverage data

Description

project new sample into the burnin space solving projection by Accelerated Proximal Gradient

Usage

apg_project(m.vec, U, lambda1, lambda2)

Arguments

m.vec	numeric vector of length m. Vector of GC corrected coverage data of sample in question
U	(m markers x n samples) numeric matrix. The basis of low rank subspace. The dimensions are same as burnin matrix
lambda1	integer. Tuning parameter (default = 1/(sqrt(no. of markers in input vector))
lambda2	integer. Tuning parameter (default = 1/(sqrt(no. of markers in input vector))

Value

list with s and v vectors

Author(s)

Aditya Deshpande

kcygan/dryclean documentation built on Feb. 13, 2020, 12:34 a.m.