simulation.random: Simulation function - random censoring

In lpetito/groupedCS: Simulations to supplement the paper "Misclassified Grouped Current Status Data"

Description

This function does one replication of the simulation for the paper from data generated with Weibull event times and Uniform censoring. It returns a description of the misclassification of both the individual and group tests, the results from the appropriate PAVA, the results from the hybrid EM-PAV algorithm for grouped tests, and the number of iterations the hybrid EM-PAV algorithm takes to converge

Usage

simulation(n, k, shape, scale, quantile, x, alpha=1, beta=1, t=0.01)

Arguments

n	number of individuals
k	grouping size
shape	shape for the Weibull distribution
scale	scale for the Weibull distribution (defaults to 1)
quantile	The maximum probability of the event in the population (default is 0.99)
x	The maximum value for censoring. This should correspond with the quantile argument.
alpha	Sensitivity: probability of a positive test results given that the individual is truly diseased (or that the group contains at least one person who is truly diseased). Default is 1 - no misclassification
beta	Specificity: probability of a negative test results given that the individual is truly not diseased (or that the group contains noone who is truly diseased). Default is 1 - no misclassification
t	threshold for convergence (default is 0.01)

Value

desc.ind	Table with description of the misclassification of the individual tests
desc.group	Table with description of the misclassification of the group tests
num.it	Number of iterations for the hybrid EM-PAV algorithm to converge
ind.result	Result from appropriate PAV algorithm (pava.cs if alpha = beta = 1, pava.cs.mc otherwise)
group.result	Result from hybrid EM-PAV algorithm, see function hybrid.em.pav for details

Author(s)

Lucia Petito

See Also

hybrid.em.pav, pava.cs.mc, gen.data.weibull.unif

Examples

simulation(100, 2, 4, 25, 0.25, 14, 0.95, 0.95, 0.01)

lpetito/groupedCS documentation built on May 21, 2019, 7:51 a.m.