mixture.sim: Simulation Of A Mixture Of Two Normal Or Gamma Distributions

In charlottedion/mixedsde: Estimation Methods for Stochastic Differential Mixed Effects Models

Description

Simulation of M random variables from a mixture of two Gaussian or Gamma distributions

Usage

mixture.sim(M, density.phi, param)

Arguments

M	number of simulated variables
density.phi	name of the chosen density 'mixture.normal' or 'mixture.gamma'
param	vector of parameters with the proportion of mixture of the two distributions and means and standard-deviations of the two normal or shapes and scales of the two Gamma distribution

Details

If 'mixture.normal', the distribution is p N(μ1,σ1^2) + (1-p)N(μ2, σ2^2)

and param=c(p, μ1, σ1, μ2, σ2)

If 'mixture.gamma', the distribution is p Gamma(shape1,scale1) + (1-p)Gamma(shape2,scale2)

and param=c(p, shape1, scale1, shape2, scale2)

Value

Y	vector of simulated variables

Examples

density.phi <- 'mixture.gamma'
param <- c(0.2,1.8,0.5,5.05,1); M <- 200
gridf <- seq(0, 8, length = 200)
f <- param[1] * 1/gamma(param[2]) * (gridf)^(param[2]-1) *
           exp(-(gridf) / param[3]) / param[3]^param[2] +
           (1-param[1]) * 1/gamma(param[4]) * (gridf)^(param[4]-1) *
           exp(-(gridf) / param[5]) / param[5]^param[4]
Y <- mixture.sim(M, density.phi, param)
hist(Y)
lines(gridf, f)

charlottedion/mixedsde documentation built on May 13, 2019, 3:35 p.m.