VSOLassoBag is a variable-selection oriented LASSO bagging algorithm for biomarker development in omic-based translational research, providing a bagging LASSO framework for selecting important variables from multiple models. A main application of this package is to screen limitted number of variables that are less dependent to train dataset. Basically, this packages was initially deveploped for adjust LASSO selected results from bootstrapped sample set. Variables with the highest frequency among the several selected result were considered as stable variables for differ sample set. However, it is usually hard to determine the cutoff in terms of frequency when applyed in a real dataset. In this package, we introduced several methods, namely (1) curve elbow point detection, (2) parametrical statistical test and (3) permutation test to help determine the cut-off point for variables selection.
You can install the released version of lassoBag from CRAN with:
install.packages("LassoBag")And the development version from GitHub with:
# install.packages("devtools")
devtools::install_github("likelet/LassoBag")A detailed documentation of VSOLassoBag could be found at here
Jiaqi Liang and Chaoye Wang from SYSU
Qi Zhao from SYSUCC, [email protected]
GNU General Public License V3