MIRL: Multiple Imputation Random Lasso for Variable Selection with Missing Entries

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.

Getting started

Package details

AuthorYing Liu, Yuanjia Wang, Yang Feng, Melanie M. Wall
MaintainerYing Liu <[email protected]>
Package repositoryView on CRAN
Installation Install the latest version of this package by entering the following in R:

Try the MIRL package in your browser

Any scripts or data that you put into this service are public.

MIRL documentation built on April 11, 2018, 5:04 p.m.