simDat: Simulate data

In ecpc: Flexible Co-Data Learning for High-Dimensional Prediction

Simulate data

Description

Simulate toy data with linear or logistic response.

Usage

simDat(n, p, n2 = 20, muGrp, varGrp, indT, sigma = 1, 
  model = c("linear","logistic"), flag = FALSE)

Arguments

n	Number of samples for the training set.
p	Number of covariates.
n2	Number of independent samples for the test set.
muGrp	Prior mean for different groups.
varGrp	Prior variance for different groups.
indT	True group index of each covariate; p-dimensional vector.
sigma	Variance parameter for linear model.
model	Type of model.
flag	Should linear predictors and true response be plotted?

Value

A list with

beta	Simulated regression coefficients
Xctd	Simulated observed data for training set
Y	Simulated response data for test set
X2ctd	Simulated observed data for test set
Y2	Simulated response data for test set

Examples

n<-10
p<-30
#simulate beta from two normal distributions; beta_k ~ N(mu_k,tau^2_k)
muGrp <- c(0,0.1) #mean (mu_1,mu_2)
varGrp <- c(0.05,0.01) #variance (tau^2_1,tau^2_2)
#group number of each covariate; first half in group 1, second half in group 2
indT <- rep(c(1,2),each=15)

dataLin <- simDat(n, p, n2 = 20, muGrp, varGrp, indT, sigma = 1, model = "linear",
    flag = TRUE)
dataLog <- simDat(n, p, n2 = 20, muGrp, varGrp, indT, model = "logistic",
    flag = TRUE)

ecpc documentation built on March 7, 2023, 6:46 p.m.