PRNGODBinBeta: Pseudo-Random Number Generation of Overdispersed Correlated...

In lalondetl/PRNG:

Description

This function produces pseudo-random numbers for responses for correlated (longitudinal) data. The responses are generated as either binary (Bernoulli) data or count (Binomial), using a transformed Beta random effect to induce autocorrelation. Once the Beta-distributed random effects are generated, the values are transformed according to a logit transformation to ensure they interact with the response similarly to otherpredictors. These random transformed effects are then passed through the systematic component to a Binomial-Beta pseudo-random number generator from the VGAM package. Therefore there are two instances of random effects: in the transformed Beta random effects, and through the Binomial-Beta pseudo-random number generation. The predictors must be provided as a vector. The function returns a response vector, labeled "Outcomes".

Usage

PRNGODBinBeta(M, mvec, D, beta0, beta1, x, seed)

Arguments

M	The number of groups in the resulting data.
mvec	A vector indicating the length of each group in the data.
D	A vector of denominators for each response; a vector of 1's indicates Bernoulli data.
beta0	The true intercept in the systematic component of the model.
beta1	The true slope in the systematic component of the model.
x	A vector of predictor values.
seed	The seed for data generation.

Examples

PRNGODBinBeta()

lalondetl/PRNG documentation built on May 20, 2019, 3:06 p.m.