Fast_LORS_Tuning: Fast_LORS_Tuning

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

Description

Fast_LORS_Tuning is a function used perform parameter tuning using FastLORS instead of LORS

Usage

Fast_LORS_Tuning(Y, X, rho, lambda, Training, Validation, maxiter = 5000,
  eps = 2.2204e-16, tol = 1e-04, B = NULL, mu = NULL, L = NULL,
  omega_SOR = 1.999)

Arguments

Y	gene expression matrix
X	matrix of SNPs
rho	parameter used to enforce sparsity of B
lambda	parameter used to enforce low-rank structure of L
Training	Boolean matrix for training data
Validation	Boolean matrix for validation data
maxiter	maximum number of iterates
eps	a small constant to prevent dividing by zero when checking relative change in function values.
tol	tolerance threshold for convergence
B	an estimate for matrix of coefficients. Default is NULL.
mu	an estimate for the intercept. Default is NULL.
L	an estimate for the hidden factors Default is NULL.
omega_SOR	the value of omega to use if applying successive over-relaxation with FastLORS.

Value

B	matrix of coefficients
L	matrix of hidden factors
mu	vector of intercepts
Err	Residual sum of squares of validation data
f_vals	Objective function values
res_vec	Relative change in objective function values
iter	Number of iterations

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