new_BGrass_chain: Initialize a new BGrass chain
In BangyaoZhao/BV: Implement the BGrass model proposed by (Zhao et al. 2022).

View source: R/new_BGrass_chain.R

new_BGrass_chain

R Documentation

Initialize a new BGrass chain

Description

Initialize a new BGrass chain

Usage

new_BGrass_chain(
  A,
  V,
  X,
  L,
  G = NULL,
  nn = rep(1, nrow(A)),
  epsilon = 0.2,
  a_alpha = 0.1,
  b_alpha = 0.1,
  a_tau = 0.5,
  b_tau = 0.5,
  a_gammaG = 0.5,
  b_gammaG = 0.5,
  pi_delta = 0.5,
  n_thin = 1,
  aggregation = TRUE,
  enrichment = FALSE
)

Arguments

`A`	A 0-1 matrix with n rows and J columns (J AEs). The i-j th element of A indicates whether the j th AE is reported in report i.
`V`	A 0-1 vector of length n indicating the type pf vaccine in each report.
`X`	The covariate matrix of dim n by p + 1 (p covariates and a column of intercept)
`L`	The J by J normalized graph Laplacian matrix of AEs. You can get it using the function `getLaplacian`.
`G`	The J by K matrix indicating the groups each AE comes from. if `!enrichment`, this argument should be set to `NULL`. You can get it using the function `getG`. (It is not useful right now)
`nn`	If `A` is already aggregated, indicate the number of reports corresponding the each row of `A`. The function will assume the data is not aggregated, so `nn = rep(1, nrow(A))` by default.
`epsilon`	A hyperparameter, controlling the level of information borrowing.
`a_alpha`	A hyperparameter
`b_alpha`	A hyperparameter
`a_tau`	A hyperparameter
`b_tau`	A hyperparameter
`a_gammaG`	For enrichment, not useful right now.
`b_gammaG`	For enrichment, not useful right now.
`pi_delta`	A hyperparameter controlling delta, only useful when not doing enrichment. If no variable selection is wanted, set `pi_delta = 1`.
`n_thin`	Thinning parameter. The number of draws for one valid MCMC sample.
`aggregation`	Whether to perform aggregation. This will usually speed up the function without changing the result.
`enrichment`	Whether to perform enrichment. This is under development, do not use this.