yaqicao/TwoPhaseAccuracy: Two-phase stratified sampling and analysis for predicting binary outcomes

In this package, function "evalTwoPhase" is for calculating AUC, TPR and FPR with two-phase estimators implemented via function tps() in R package "osDesign" (Haneuse and others, 2011). Function "seTwoPhase" is for estimating the corresponding standard errors using bootstrap resampling. The input arguments of the two functions include threshold values for estimating TPR and FPR and a dataframe that contains case-control status, stratum membership, Phase I predictors, Phase II predictors, variable names and a variable indicating whether a subject was selected into Phase II. Function "seTwoPhase" includes an additional input argument for the number of bootstrap samples. A third function, "summaryTwoPhase" summarizes the results from "evalTwoPhase and "seTwoPhase by outputing the estimates of AUC, TPR, and FPR together with their standard drror estimates.

install.packages("remotes")
remotes::install_github("yaqicao/TwoPhaseAccuracy")