In kjytay/cvwrapr: Tools for Cross Validation

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.path = "man/figures/README-",
  out.width = "100%"
)

cvwrapr

The goal of cvwrapr is to make cross-validation (CV) easy. The main function in the package is kfoldcv. It performs K-fold CV for a hyperparameter, returning the CV error for a path of hyperparameter values along with other useful information. The computeError function allows the user to compute the CV error for a range of loss functions from a matrix of out-of-fold predictions.

See the package website for more examples and documentation.

Installation

You can install cvwrapr in the following ways:

# from CRAN
install.packages("cvwrapr")

# development version from GitHub
# install.packages("devtools")
devtools::install_github("kjytay/cvwrapr")

Example

This is a basic example showing how to perform cross-validation for the lambda parameter in the lasso (Tibshirani 1996).

# simulate data
set.seed(1)
nobs <- 100; nvars <- 10
x <- matrix(rnorm(nobs * nvars), nrow = nobs)
y <- rowSums(x[, 1:2]) + rnorm(nobs)

library(cvwrapr)
library(glmnet)

set.seed(1)
cv_fit <- kfoldcv(x, y, train_fun = glmnet, predict_fun = predict)

The returned output contains information on the CV procedure and can be plotted.

names(cv_fit)
plot(cv_fit)

kjytay/cvwrapr documentation built on Dec. 21, 2021, 6:45 a.m.