- Code optimization and bug fixes.
- Function naming convention update.
- Allows contrast and subset in
hima
. - Add vignette for HIMA.
- Coding optimization and bug fixes.
- Renmae
hima2
tohima
(hima2
will be retired by 2025-11-1) and improved parameter naming.
- Add efficient HIMA function
eHIMA
. - Optimization of function
hima2
. - Stablization of estimates in function
qHIMA
. - Some improvements in output presentation and consistency.
- Add quantile HIMA function
qHIMA
. - Enable de-biased lasso penalty for regular HIMA in function
hima2
- Add wrapper function
hima2
for one-stop solution of different HIMA analysis.
- Rename
family
option inhima
function toY.family
. - Rename
screen.family
option inhima
function toM.family
- Improve
verbose
messages.
- Keep older version of
null_estimation
in packageHDMT
(< 1.0.4). - Allows negative binomial regression in funcion
hima
when outcome is binary. - Add new function
microHIMA
to handle compositional microbiome data.
- Add new function
survHIMA
to handle survival data.
- Allow low-dimensional mediator selections.
- Two different sets of covariates are allowed for exposure and outcome separately.
- Benjamini-Hochberg FDR adjustment is available.
- Added error handler when no mediators were selected from penalized model.
- More transparent information returned during model fitting (when
verbose = TRUE
).
- Improvement: More powerful model fit when outcome variable is binary.
- Generalized linear regression is now allowed in screening step.
- Initial package submission to GitHub