Implements a variable selection and prediction method for high-dimensional data with missing entries following the paper Liu et al. (2016) <doi:10.1214/15-AOAS899>. It deals with missingness by multiple imputation and produces a selection probability for each variable following stability selection. The user can further choose a threshold for the selection probability to select a final set of variables. The threshold can be picked by cross validation or the user can define a practical threshold for selection probability. If you find this work useful for your application, please cite the method paper.
| Version: | 1.0 |
| Depends: | glmnet, mice, MASS, boot |
| Published: | 2018-04-11 |
| Author: | Ying Liu, Yuanjia Wang, Yang Feng, Melanie M. Wall |
| Maintainer: | Ying Liu <summeryingl at gmail.com> |
| License: | GPL-2 |
| NeedsCompilation: | no |
| CRAN checks: | MIRL results |
| Reference manual: | MIRL.pdf |
| Package source: | MIRL_1.0.tar.gz |
| Windows binaries: | r-devel: MIRL_1.0.zip, r-release: MIRL_1.0.zip, r-oldrel: MIRL_1.0.zip |
| OS X binaries: | r-release: MIRL_1.0.tgz, r-oldrel: MIRL_1.0.tgz |
Please use the canonical form https://CRAN.R-project.org/package=MIRL to link to this page.