ROC: ROC curve analysis

In Dingding-Hu/BPLR-package: ROC curve analysis

ROC curve analysis

Description

To estimate the ROC curve and its summary statistics including the AUC, Youden index, and optimal cutoff point based on a single biomarker by eight differet methods. The default is the Bernstein polynomial approach in Hu et al. (2022), which is developed under the likelihood ratio ordering assumption.

Usage

ROC(x, y, method = "BP", nss = 10^4)

Arguments

x	vector contains the sample of biomarkers from the "healthy" group
y	vector contains the sample of biomarkers from the "diseased" group
method	the method to estimate ROC curve and its summary statistics. It can be "BP","Box-Cox","ZL","ECDF","LZL","MNLE","Kernel",or "MSLE".
nss	control the number of point estimates of the ROC curve in the range [0,1], default is 10^4.

Details

The detail for each method is provided in Hu et al. (2022) or https://github.com/Dingding-Hu/BPLR-package.

Examples

x=rnorm(100,10,1)
y=rnrom(100,12,1)
ROC(x,y,method="BP")

Dingding-Hu/BPLR-package documentation built on March 26, 2023, 11:43 p.m.