computeAUC: computeAUC: computes AUC

In RadioGx: Analysis of Large-Scale Radio-Genomic Data

Description

This function computes the area under a dose-response curve of the form survival fraction SF = exp(-alpha * D - beta * D ^ 2).

Usage

computeAUC(
  D,
  SF,
  pars,
  lower,
  upper,
  trunc = TRUE,
  SF_as_log = FALSE,
  area.type = c("Fitted", "Actual"),
  verbose = TRUE
)

Arguments

D	vector of dosages
SF	vector of survival fractions
pars	parameters (alpha, beta) in equation y = exp(-alpha * x - beta * x ^ 2)
lower	lower bound of dose region to compute AUC over
upper	upper bound of dose region to compute AUC over
trunc	should survival fractions be truncated downward to 1 if they exceed 1?
SF_as_log	A boolean indicating whether survival fraction is displayed on a log axis. Defaults to FALSE
area.type	should the AUC of the raw (D, SF) points be returned, or should the AUC of a curve fit to said points be returned instead?
verbose	how detailed should error and warning messages be? See details.

Details

If lower and/or upper are missing, the function assumes their values to be the minimum and maximum D-values, respectively. For all warnings to be silent, set trunc = FALSE. For warnings to be output, set trunc = TRUE. For warnings to be output along with the arguments that triggered them, set trunc = 2.

Value

numeric The area under the ROC curve

Examples

computeAUC(D=c(0.1, 0.5, 0.7, 0.9), pars=c(0.2, 0.1), lower = 0,
 upper = 1) # Returns 0.7039296

RadioGx documentation built on Nov. 8, 2020, 8:21 p.m.