COPY_biglasso_part: Train one model

In bigstatsr: Statistical Tools for Filebacked Big Matrices

Train one model

Description

Train one model

Usage

COPY_biglasso_part(
  X,
  y.train,
  ind.train,
  ind.col,
  covar.train,
  family,
  lambda,
  center,
  scale,
  resid,
  alpha,
  eps,
  max.iter,
  dfmax,
  ind.val,
  covar.val,
  y.val,
  n.abort,
  nlam.min,
  base.train,
  base.val,
  pf
)

Value

A named list with following variables:

intercept	A vector of intercepts, corresponding to each lambda.
beta	The vector of coefficients that minimized the loss on the validation set.
iter	A vector of length nlambda containing the number of iterations until convergence at each value of lambda.
lambda	The sequence of regularization parameter values in the path.
alpha	Input parameter.
loss	A vector containing either the residual sum of squares (for linear models) or negative log-likelihood (for logistic models) of the fitted model at each value of lambda.
loss.val	A vector containing the loss for the corresponding validation set.
message	Reason the fitting has stopped.
nb_active	The number of active (non-zero) variables along the regularization path.
nb_candidate	The number of candidate variables (used in the gradient descent) along the regularization path.
ind.train	Indices of training set.

bigstatsr documentation built on Sept. 11, 2024, 7:08 p.m.