couple_metropolis_to_condbern: Coupled Metropolis algorithm for the Conditional-Bernoulli...

In nianqiaoju/agents: Toolkit for inference in agent-based modelss

Description

A coupling of two metropolis samplers targetting the same Conditional-Bernoulli distribution Chain1 starts from the stationary distribution and chain2 is initialized uniformly from the support

Usage

couple_metropolis_to_condbern(
  alpha,
  sum_x,
  num_mcmc,
  compute_sse = FALSE,
  verbose = FALSE
)

Arguments

alpha	vector of probabilities for the Poisson-Binomial distribution
sum_x	sum of x
num_mcmc	number of MCMC iterations; if num.mcmc = Inf, then run the algorithm until the two chains couple
compute_sse	binary, default is FALSE; if compute_sse = TRUE, then returns the sum-squared-error distance from chain1 to chain2
verbose	binary, whether to print messages

Value

a list that contains

• meeting time

• sum-squared-error distance from chain1 to chain2

nianqiaoju/agents documentation built on Feb. 19, 2021, 12:18 a.m.