posterior_prob_gamma_poisson {oHMMed}

Forward-Backward Algorithm to Calculate the Posterior Probabilities of Hidden States in Poisson-Gamma Model

Description

Forward-Backward Algorithm to Calculate the Posterior Probabilities of Hidden States in Poisson-Gamma Model

Usage

posterior_prob_gamma_poisson(data, pi, mat_T, betas, alpha)

Arguments

data

(numeric) Poisson data

pi

(numeric) prior probability of states

mat_T

(matrix) transition probability matrix

betas

(numeric) vector with prior rates

alpha

(numeric) prior scale

Details

Please see supplementary information at doi:10.1186/s12859-024-05751-4 for more details on the algorithm.

Value

List with the following elements:

• F: auxiliary forward variables

• B: auxiliary backward variables

• s: weights

Examples

mat_T <- rbind(c(1-0.01,0.01,0),
               c(0.01,1-0.02,0.01),
               c(0,0.01,1-0.01))
L <- 2^10
betas <- c(0.1, 0.3, 0.5)
alpha <- 1

sim_data <- hmm_simulate_gamma_poisson_data(L = L,
                                            mat_T = mat_T,
                                            betas = betas,
                                            alpha = alpha)
pi <- sim_data$pi
hmm_poison_data <- sim_data$data
hist(hmm_poison_data, breaks = 100)

# Calculate posterior probabilities of hidden states
post_prob <- posterior_prob_gamma_poisson(data = hmm_poison_data,
                                          pi = pi,
                                          mat_T = mat_T,
                                          betas = betas,
                                          alpha = alpha)
str(post_prob)