SIM.FUN: Generate simulation data to test adaptive lasso

In michaelyanwang/divideconquer: Divide-and-conquer adaptive lasso using least square approximation

Description

SIM.FUN generates binary, count, and continuous-time survival response data that are associated with design matrix. The design matrix comes from a correlated multivariate normal. The default signals (beta0) are sparse.

Usage

SIM.FUN(nn, p.x = 50, cor = 0.2, family = c("binary", "count", "Cox"),
  beta0 = NULL)

Arguments

nn	sample size
p.x	number of covariates
cor	correlation of covariates
family	the family of response data taking c('binary','count','Cox')
beta0	the coefficients for the design, including intercept

Value

For binary and count data, it returns a matrix with the first column=response, rest = design matrix, without intercept. For survival data, it returns a matrix with the first column U, second column delta (0,1), and rest = design matrix.

Author(s)

Yan Wang, Tianxi Cai

Examples

SIM.FUN(nn = 1e6, p.x = 50, family = 'binary')

michaelyanwang/divideconquer documentation built on Aug. 16, 2019, 10:11 a.m.