In david26694/bolasso: Bolasso

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.path = "README-"
)

bolasso

The goal of bolasso is to apply robust feature selection via lasso and bootsrtapping.

Installation

You can install bolasso from github with:

# install.packages("devtools")
devtools::install_github("david26694/bolasso")

Example

This is a basic example which shows you how to apply bolasso in a regression setting:

## basic example code
library(bolasso)

# We want to predict vs with the other mtcars variables
predictors <- mtcars[,c("mpg", "disp", "hp", "drat", "wt")]
outcome <- mtcars[, "vs"]

# Apply bolasso with 10 bootstrap samples
model <- bolasso(predictors, outcome, n_bootstraps = 10)

# Show variable selection results
summary(model)

# It can also run on recipes, and with formulas
library(recipes)
rec <- recipe(mpg ~ ., mtcars)
rec <- step_log(rec, disp)
model_rec <- bolasso(rec, mtcars, n_bootstraps = 20)

summary(model_rec)

# It has different types of prediction, like numeric, class and prob
predict(model_rec, mtcars, type = 'numeric')