ridge_solver_interaction: Computes the ridge regularized estimates of the interaction...

In goepp/hazreg: Hazard Regularized estimation for time-to-event data

Description

Computes the ridge regularized estimates of the interaction model

Usage

ridge_solver_interaction(O, R, pen, weights_age = NULL,
  weights_cohort = NULL, use_band = TRUE, old_par = rep(0, ncol(O) *
  nrow(O)), maxiter = 1000, verbose = FALSE, tol = 1e-08)

aridge_solver_interaction(O, R, pen, sample_size, use_band = TRUE,
  maxiter = 1000, verbose = FALSE, progress = TRUE)

Arguments

O	Observed events as returned by exhaustive_stat_2d
R	Time at risk as returned by exhaustive_stat_2d
pen	The penalty constant.
weights_age	weights of the age differences in the log-hazard. See ADD REFERENCE SHEET
weights_cohort	weights of the cohort differences in the log-hazard. See ADD REFERENCE SHEET
use_band	Whether to use faster inversion of the hessian function. TRUE
old_par	initial point of the iteration. The default choice is 0
maxiter	The maximal number of iteractions of the Newton-Raphson procedure.
verbose	Whether to display the progress of the Newton-Raphson procedure.
tol	Absolute error tolerance used as stopping criteria of the Newton-Raphson minimization. is recommended

Value

The vector estimate of the ridge regularized interaction model.

See Also

Other ridge: ridge_solver_aci

Other adaptive_ridge: ridge_solver_aci

Examples

## Not run: 
J <- 10
K <- 15
set.seed(0)
O <- matrix(rpois(K * J, 2), K, J)
R <- matrix(rpois(K * J, 10), K, J)
pen <-  50
ridge <- ridge_solver_interaction(O, R, pen)
ridge$par
ridge$iter

## End(Not run)

goepp/hazreg documentation built on June 7, 2019, 4:03 a.m.