sampler manifestos 0.8

Control the Generation of Random Numbers for Sampling from the Proposal Distribution

Description

The functions mvt.sampler and mvnorm.sampler can be used to fine-tune the construction and sampling from the proposal distribution in the importance sampling algorithm used in the MCEM algorithm to fit latpos models.

Usage

mvt.sampler(df,fix.seed=TRUE) # Multivariate Student's t-distribution
mvnorm.sampler(fix.seed=TRUE) # Multivariate normal distribution

Arguments

df

a number, the degrees of freedom of the multivariate Student’s t-distribution.

fix.seed

a logical value; if TRUE the seed of the random number generated is reset in each MCEM iteration to make sure that approximate marginal log-likelihood surface is continuous while the simulation sample size is constant.

Value

Both functions return a list that contains a function to generate random numbers and to compute the log-density function of the distribution from which random numbers are generated.