generate_pop_clust_strat: Generate a stratified population of clustered observations

In tamytsujimoto/surveyROC: ROC curve for survey data

Description

Generate a population containing clusters containing healthy and diseased observations generated from normal distributions. The correlated binary data is generated using the multivariate probit method. The correlated continuous marker values are generated by including a random effect

Usage

generate_pop_clust_strat(
  N,
  freq_strata,
  mean0_strat,
  mean1_strat,
  sigma0_strat,
  sigma1_strat,
  p1,
  cluster_size,
  tau
)

Arguments

mean0_strat	numeric vector of mean for class label 0 (healthy) for all strata
mean1_strat	numeric vector of mean for class label 1 (diseased) for all strata
sigma0_strat	numeric vector of standard deviation for class label 0 (healthy) for all strata
sigma1_strat	numeric vector of standard deviation for class label 1 (diseased) for all strata
p1	numeric value for proportion of diseased observations in the population
cluster_size	list of cluster sizes per strata
corr_mvn	numeric value for correlation in multivariate normal distribution

Value

Dataframe containing

• cluster_size: cluster size

• X: continuous marker values

• D: class labels (0: healthy; 1: diseased)

tamytsujimoto/surveyROC documentation built on Sept. 8, 2021, 11:35 p.m.