Fast_LORS: Fast_LORS

In jdrhyne2/FastLORS: Joint Modeling for eQTL Mapping in R

Description

Fast_LORS is a function for solving the LORS optimization problem in Can Yang et al. (2013) through the proximal gradient method

Usage

Fast_LORS(Y, X, rho, lambda, maxiter = 5000, eps = 2.2204e-16,
  tol = 1e-04, verbose = FALSE, omega_SOR = 1.999)

Arguments

Y	gene expression matrix
X	matrix of SNPs
rho	parameter for enforcing sparsity of coefficient matrix
lambda	parameter for enforcing low-rank structure of hidden factor matrix
maxiter	maximum number of iterations
eps	constant used when checking the convergence. Ensures no division by 0.
tol	tolerance level for convergence
verbose	chooses whether details should be printed to console. Default is FALSE.
omega_SOR	the value of omega to use if applying successive over-relaxation with FastLORS.

Value

B	The estimated coefficients
mu	The estimated intercept
L	The estimated matrix of hidden factors
f_val_vec	The objective function values
res_vec	The relative change in objective function values
iter	The number of iterations

Examples

##Example

## Generate some data
n <- 50
p <- 200
q <- 100
k <- 10
set.seed(123)
X <- matrix(rbinom(n*p,1,0.5),n,p)
L <- matrix(rnorm(n*k),n,k) %*% t(matrix(rnorm(q*k),q,k))
B <- matrix(0, ncol(X), ncol(L))
activeSNPs <- sort(sample(c(1:nrow(B)), 20))
for(i in 1:length(activeSNPs)){
genes_influenced <- sort(sample(c(1:ncol(B)),5))
B[activeSNPs[i], genes_influenced] <- 2
}
E <- matrix(rnorm(n*q),n,q)
Y <- X %*% B + L + E

rho <- runif(1,3,5)
lambda <- runif(1,3,5)

## Usage
Fast_LORS(Y, X, rho, lambda)

jdrhyne2/FastLORS documentation built on March 5, 2020, 6:50 a.m.