In benkeser/nlpred: Estimators of Non-Linear Cross-Validated Risks Optimized for Small Samples

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

R/nlpred

Small-sample optimized estimators of cross-validated prediction metrics

David Benkeser

---

Description

nlpred is an R package for computing estimates of cross-validated prediction metrics. These estimates are tailored for superior performance in small samples. Several estimators are available including ones based cross-validated targeted minimum loss-based estimation, estimating equations, and one-step estimation.

---

Installation

For standard use, we recommend installing the package from CRAN via

install.packages("nlpred")

You can install the current release of nlpred from GitHub via devtools with:

devtools::install_github("benkeser/nlpred")

---

Usage

The main functions in the package are cv_auc and cv_scrnp, which are used to compute, respectively, the K-fold cross-validated area under the receiver operating characteristics curve (CVAUC) and the K-fold cross-validated sensitivity constrained rate of negative prediction. However, rather than using standard cross-validation estimators (where prediction algorithms are developed in a training sample and AUC/SCRNP estimated using the validation sample), we instead use techniques from efficiency theory to estimate these quantities. This allows us to use the training data both to develop the prediction algorithm, as well as key nuisance parameters needed to evaluate AUC/SCRNP. By reserving more data for estimation of these key parameters, we obtain improved performance in small samples.

# load package
library(nlpred)

# turn off messages from np package
options(np.messages=FALSE)

# simulate data
n <- 200
p <- 10
X <- data.frame(matrix(rnorm(n*p), nrow = n, ncol = p))
Y <- rbinom(n, 1, plogis(X[,1] + X[,10]))

# get cv auc estimates for logistic regression
logistic_cv_auc_ests <- cv_auc(Y = Y, X = X, K = 5, learner = "glm_wrapper")
logistic_cv_auc_ests

# get cv auc estimates for random forest using nested 
# cross-validation for nuisance parameter estimation. nested
# cross-validation is unfortunately necessary when aggressive learners 
# are used. 
rf_cv_auc_ests <- cv_auc(Y = Y, X = X, K = 5, 
                         learner = "randomforest_wrapper", 
                         nested_cv = TRUE)
rf_cv_auc_ests

# same examples for scrnp
logistic_cv_scrnp_ests <- cv_scrnp(Y = Y, X = X, K = 5, learner = "glm_wrapper")
logistic_cv_scrnp_ests


rf_cv_scrnp_ests <- cv_scrnp(Y = Y, X = X, K = 5, 
                         learner = "randomforest_wrapper", 
                         nested_cv = TRUE)
rf_cv_scrnp_ests

---

Issues

If you encounter any bugs or have any specific feature requests, please file an issue.

---

Contributions

Interested contributors can consult our contribution guidelines prior to submitting a pull request.

---

Citation

After using the nlpred package, please cite the following:

@Manual{nlpredpackage,
  title = {nlpred: Estimators of Non-Linear Cross-Validated Risks Optimized for Small Samples},
  author = {David Benkeser},
  note = {R package version 1.0.1}
}

@article{benkeser2019improved,
  year  = {2019},
  author = {Benkeser, David C and Petersen, Maya and van der Laan, Mark J},
  title = {Improved Small-Sample Estimation of Nonlinear Cross-Validated Prediction Metrics},
  journal = {Journal of the American Statistical Association},
  doi = {10.1080/01621459.2019.1668794}
}

License

© 2019- David Benkeser

The contents of this repository are distributed under the MIT license. See below for details:

The MIT License (MIT)

Copyright (c) 2019- David C. Benkeser

Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.

benkeser/nlpred documentation built on June 11, 2022, 3:05 a.m.