HZINB_one_gamma: HGZIPS - HZINB (not assuming independence)

In sidiwang/hgzips: hgzips

Description

This HZINB_one_gamma function finds hyperparameter estimates by implementing the Expectation-Maximization (EM) algorithm and hierarchical zero-inflated negative binomial model with one gamma component.

Usage

parRangeCheck(N_ij, E_ij)

grid_HZINB(a_j, b_j, omega_j, K, L, H)

HZINB_one_gamma(
  grid_a,
  grid_b,
  grid_omega,
  init_pi_klh,
  dataset,
  iteration,
  Loglik = FALSE,
  zeroes = FALSE,
  N_star = 1
)

Arguments

N_ij	matrix of N_ij values, i - AE, j - drugs
E_ij	matrix of E_ij values, corresponding to N_ij
a_j	alpha for gamma distribution
b_j	beta for gamma distribution
omega_j	proportion
K	number of a_j in grid
L	number of b_j in grid
H	number of omega_j in grid
grid_a	alpha value grid
grid_b	beta value grid
grid_omega	omega value grid
init_pi_klh	initial probability value of all the alpha, beta, omega combinations for implementing the EM algprithm
dataset	a list of squashed/unsquashed datasets that include N_ij, E_ij and weights for each drug (j). This dataset list can be generated by the rawProcessing function in this package.
iteration	number of EM algorithm iterations to run
Loglik	whether to return the loglikelihood of each iteration or not (TRUE or FALSE)
zeroes	A logical scalar specifying if zero counts should be included.
N_star	the minimum Nij count size to be used for hyperparameter estimation. If zeroes are included in Nij vector, please set N_star = NULL

Value

parRangeCheck the estimated a_j, b_j and omega_j for each drug (j) parRangeCheck

grid_HZINB build a suitable grid of a_j, b_j, and omega_j for implementing HZINB grid_HZINB

HZINB_one_gamma a list of estimated probability of each alpha, beta, omega combination and their corresponding loglikelihood (optional)

• theta_EM Estimate of hyperparameters for each EM iteration

• llh logliklihood for each EM iteration (optional)

HZINB_one_gamma

sidiwang/hgzips documentation built on Jan. 19, 2021, 4:09 p.m.