parallel_sequential_fusion_mixG: Standard Sequential Monte Carlo Fusion

In rchan26/mixGaussTempering: What the Package Does (One Line, Title Case)

Description

Sequential Monte Carlo Fusion with base level with nodes that are tempered mixture Gaussians with same weights, means, sds components

Usage

parallel_sequential_fusion_mixG(N_schedule, time_schedule, start_beta,
  base_samples, weights, means, sds, study = F)

Arguments

N_schedule	vector of length (L-1), where N_schedule[l] is the number of samples per node at level l
time_schedule	vector of legnth(L-1), where time_schedule[l] is the time chosen for Fusion at level l
start_beta	beta for the base level
base_samples	list of length (1/start_beta), where samples_to_fuse[c] containg the samples for the c-th node in the level
weights	vector: weights of mixture Gaussian
means	vector: means of mixture Gassuan
sds	vector: st.devs of mixture Gaussian
study	boolean value: defaults to F, determines whether or not to return acceptance probabilities

Value

samples: samples from Sequential fusion

time: vector of length (L-1), where time[l] is the run time for level l

rho_acc: vector of length (L-1), where rho_acc[l] is the acceptance rate for first fusion step for level l

Q_acc: vector of length (L-1), where Q_acc[l] is the acceptance rate for second fusion step for level l

input_betas: list of length (L), where input_betas[[l]] is the input betas for level l

output_beta: vector of length(L-1), where output_beta[l] is the beta for level l

diffusion_times: list of length (L-1), where diffusion_times[[l]] are the scaled/adapted times for fusion in level l

Examples

# setting variables
w_example <- c(0.35, 0.65)
m_example <- c(-3, 8)
s_example <- c(1, 1.5)
b_example <- 1/4

# sampling from tempered density
nsamples <- 500000
base <- hmc_base_sampler_mixG(w_example, m_example, s_example, b_example, nsamples, 4)

test <- parallel_sequential_fusion_mixG(N_schedule = rep(10000, 3), time_schedule = rep(1, 3),
                                        start_beta = b_example, base_samples = base,
                                        weights = w_example, means = m_example, sds = s_example, study = T)

# plot results
plot_levels_seq_mixG(test, weights = w_example, means = m_example, sds = s_example,
                     from = -15, to = 20, plot_rows = 2, plot_columns = 2, bw = c(0.2, 0.3, 0.3, 0.35))

rchan26/mixGaussTempering documentation built on June 14, 2019, 3:26 p.m.