CVWeightedBootstrap: CVWeightedBootstrap
In wlktan/surrogateSampling: Functions to conduct surrogate guided sampling designs

Runs a K-fold cross-validation with weighted bootstrap to approximate training/testing error from an annotated dataset

1 2	CVWeightedBootstrap(feature.mat, k_fold = 5, boot_samples = 100, pz, stratify_var = quo(Z), Y = "Y", model = "lasso")

`feature.mat`	Data frame containing features and labels
`k_fold`	Number of folds for cross validation
`boot_samples`	Number of weighted bootstrap samples for validation
`pz`	Prevalence of surrogate positives in the cohort i.e. P(Z=1)
`stratify_var`	Surrogate name in quo(.)
`Y`	Column name of the binary outcome
`model`	Whether to use a glm, lasso, or ridge model

A data frame with k_fold rows, where each row has the AUC's obtained from training, validation (naive, no reweighting), as well as the mean and variance of the weighted bootstrap resampling

wlktan/surrogateSampling documentation built on May 23, 2019, 3:07 p.m.