Implementation of the SuSiF method
Usage
susiF.workhorse(
obj,
W,
X,
tol,
init_pi0_w,
control_mixsqp,
indx_lst,
lowc_wc,
nullweight,
cal_obj,
verbose,
cov_lev,
min_purity,
maxit,
tt,
parallel = FALSE,
max_SNP_EM = 500,
max_step_EM = 1,
cor_small = FALSE,
is.pois = FALSE,
e = 0.001
)Arguments
- obj
an object of class susiF
- W
a list in which element D contains matrix of wavelet d coefficients and element C contains the vector of scaling coefficients
- X
matrix of size n by p contains the covariates
- tol
a small, non-negative number specifying the convergence tolerance for the IBSS fitting procedure. The fitting procedure will halt when the difference in the variational lower bound, or “ELBO” (the objective function to be maximized), is less than
tol.- init_pi0_w
starting value of weight on null compoenent in mixsqp (between 0 and 1)
- control_mixsqp
list of parameter for mixsqp function see mixsqp package
- indx_lst
list generated by gen_wavelet_indx for the given level of resolution
- lowc_wc
list of wavelet coefficients that exhibit too little variance
- nullweight
numeric value for penalizing likelihood at point mass 0 (should be between 0 and 1) (usefull in small sample size)
- cal_obj
logical if set as TRUE compute ELBO for convergence monitoring
- verbose
If
verbose = TRUE, the algorithm's progress, and a summary of the optimization settings are printed to the console.- cov_lev
numeric between 0 and 1, corresponding to the expected level of coverage of the cs if not specified set to 0.95
- min_purity
minimum purity for estimated credible sets
- maxit
Maximum number of IBSS iterations.
- tt
output of the cal_Bhat_Shat function, if provided allow skipping the first step of the IBSS
- parallel
if true use parallel computation
- max_SNP_EM
check susiF description
- max_step_EM
max_step_EM
- cor_small
check susiF description
- is.pois
check susiF description
- e
threshold value to avoid computing posterior that have low alpha value. Set it to 0 to compute the entire posterio. default value is 0.001